Tcp/ip packet-centric wireless transmission system architecture

Description: TECHNICAL FIELD [Detailed description of the invention]

【0001】

(Background of invention) Cross-reference with other patents [0002]

The following applications of the co-assignee include common disclosures. A US patent application filed on July 9, 1999, entitled "Quality of Service (QoS) -Aware Wireless Point to Multi-Point (PtMP) Transmission System Architecture". Agent reference number A-21506. A US patent application entitled "Method for Providing Dynamic Bandwidth Allocation Based on IP-Flow Characteristics in a Wireless Point to Multi-Point (PtMP) Transmission System" filed on July 9, 1999. Agent reference number A-21507. "Method for Providing for Quality of Service (QoS) _ Based Handling of IP-Flows in a Wireless Point to Multi-Point Transmission A US patent application entitled "System" filed on July 9, 1999. Agent reference number A-21508. A US patent application filed on July 9, 1999 entitled "IP-Flow Identification in a Wireless point to Multi-Pont Transmission System". Agent reference number A-21509. A US patent application filed on July 9, 1999, entitled "IP-Flow characterization in a Wireless Point to Muti-Point (PtMP) Transmission System". Agent reference number A-21510. A US patent application filed on July 9, 1999, entitled "IP-Flow Classification in a Wireless Point to Multi-Point (PtMP) Transmission System". Agent reference number A-21511. "IP-Flow Prioritization in a Wireless Point to Multi-Point (PtMP) A US patent application filed on July 9, 1999, entitled "Transmission System." Agent reference number A-21512. A US patent application filed on July 9, 1999, entitled "Method of Operation for Providing for Service Level Agreements (SLA) Based Prioritization in a Wireless Point to Multi-Point (PtMP) Transmission System." Agent reference number A-21513. A US patent application filed on July 9, 1999, entitled "Method for Transmission Control Protocol (TCP) Rate Control With Link-Layer Acknowledgments in a Wireless Point to Multi-Point (PtMP) Transmission System." Agent reference number A-21514. "Transmission Control Protocol / Internet Protocol (TCP / IP) _ Centric QoS Aware Media Access Control (MAC) Layer in a Wireless Point to Muti-Point (PtMP) Transmission System "A US patent application filed on July 9, 1999. Agent reference number A-21532. A US patent application filed on July 9, 1999, entitled "Use of Priority-Based Scheduling for the Optimization of Latency and Jitter Sensitive IP Flows in a Wireless Point to Multi-Point Transmission System." Agent reference number A-21533. "Time Division Multiple Access / Time Division Duplex (TDMA / TDD) Access Method for a Wireless Point to Multi-Point Transmission A US patent application entitled "System" filed on July 9, 1999. Agent reference number A-21534. A US patent application entitled "Reservation Based Prioritization Method for Wireless Transmission of Latency and Jitter Sensitive IP-Flows in a Wireless Point to Multi-Point Transmission System" filed on July 9, 1999. Agent reference number A-21535. A US patent application filed on July 9, 1999, entitled "Translation of Internet-Prioritized Internet Protocol (IP) _ Flows into Wireless System Resource Allocations in a Wireless Point to Multi-Point (PtMP) Transmission System". Agent reference number A-21536. "Method of Operation for the A US patent filed on July 9, 1999 entitled "Integration of Differentiated services (Diff-serv) Marked IP-Flows into a Quality of Service (QoS) Priorities in a Wireless Point to Multi-Point (PtMP) Transmission System" application. Agent reference number A-21539. A US patent application filed on July 9, 1999, entitled "Method for the Recognition and Operation of Virtual Private networks (VPNs) over a Wireless Point to Multi-Point (MtMP) Transmission System." Agent reference number A-21540. "Time Division Multiple Access / Time Division Duplex (TDMA / TDD) Transmission media Access Control (MAC) A U.S. patent application filed on July 9, 1999, entitled "Aire Frame." Agent reference number A-21541. "Application-Aware, Quality of Service (QoS) Sensitive, A US patent application filed on July 9, 1999, entitled "Media Access Control (MAC) Layer". Agent reference number A-21542. A US patent application filed on July 9, 1999 entitled "Transmission Control Protocol / Internet Protocol (TCP / IP) Packet-Centric Wireless Point to Point (PtP) Transmission System Architecture". Agent reference number A-21543. A US patent application filed on July 9, 1999 entitled "Transmission Control Protocol / Internet Protocol (TCP / IP) Packet-Centric Cable Point-to Multi-Point (PtMP) Transmission System Architecture". Agent reference number A-21547. [0003]

(Technical field to which the invention belongs) [0004]

The present invention relates generally to telecommunications, and more specifically to systems and methods for implementing quality of service (QoS) sensitive point-to-multipoint transmission systems. [0005]

(Conventional technology) [0006]

Telecommunications networks, such as voice, data and video networks, have traditionally been customized for each type of traffic they send. For example, voice traffic is very latency sensitive, but quality is not so important, so voice networks are designed to send voice traffic with limited latency. In contrast, traditional data traffic, such as spreadsheets, is not latency sensitive, but is required to be transmitted accurately. Traditional telecommunications networks perform circuit switching to achieve acceptable quality of service (QoS) for end users. With the advent of new packet-switched high-bandwidth data networks, it has become possible to carry different types of traffic over data networks. In particular, it has become possible to consolidate voice, data and video networks into a single broadband telecommunications network. In order to meet the wishes of end users, there is a need for a system that can provide QoS for various types of traffic to be transmitted. [0007]

(Problems to be solved by the invention) [0008]

Currently, wireless networks have a problem in QoS as compared with priority networks. For example, conventional wireless networks have a high BER (bit error rate) for several reasons. Traditionally, wireless networks also perform circuit-switched connections to provide reliable communication channels. However, circuit-switched connections allocate bandwidth between communication nodes, whether or not traffic is being transmitted between the nodes. Therefore, circuit-switched connections do not use communication bandwidth efficiently. [0009]

Packet switching makes better use of available bandwidth than traditional circuit switching. Packet switching can divide traffic into so-called "packets", transmit them from the source node to the destination, and reassemble them. Therefore, a specific part of the bandwidth can be shared by many sources and destinations, and the bandwidth can be used more effectively. [0010]

There is a demand for a wireless broadband access telecommunications system that can provide a QoS function equivalent to the transmission of a wired broadband access device. While providing sufficient bandwidth for a wide band, the lack of acceptable QoS characteristics has traditionally been a barrier to the deployment of wireless wideband access systems. Similarly, providing high levels of QoS at the expense of sufficient bandwidth was also not beneficial to the end user. [0011]

Efforts to provide conventional wireless broadband access systems have not been optimally designed because QoS has not been given sufficient priority as a basic principle of wireless system architecture. The rapid rise of the Internet, the packet-switched paradigm, and the Transmission Control Protocol / Internet Protocol (TCP / IP) as a universal data protocol has made it clear that new wireless system designs are needed. [0012]

There is a need for an IP-centric wireless broadband access system with true QoS capabilities. [0013]

(Means to solve the problem) [0014]

The present invention is via a radio base station that communicates with a first data network using a packet-centric protocol, one or more host workstations that communicate with the first data network using the packet-centric protocol, and a wireless medium. One or more customer premises equipment (CPE) stations connected to the radio base station via a shared bandwidth by the packet-centric protocol and the subscriber CPE via a second network. It relates to a packet-centric radio point-to-multipoint telecommunications system consisting of one or more subscriber workstations connected to each of the stations by the packet-centric protocol. The packet-centric protocol may be a transmission control protocol / Internet protocol (TCP / IP). The packet-centric protocol may be a user diagram protocol / internet protocol (UDP / IP). [0015]

A resource allocation means for allocating shared bandwidth between the subscriber CPE stations may be provided. The resource allocation means can optimize the quality of service (QoS) of the final user. The wireless communication medium can be composed of one or more of a radio frequency (RF) communication medium; a cable communication medium; and a satellite communication medium ;. The wireless communication medium is a time division multiple access (TDMA) access method; a time division multiple access / time division duplex (TDMA / TDD) access method; a code division multiple access (CDMA) access method; and a frequency division multiple access ( It can include one or more of FDMA) access methods; [0016]

The first data network can consist of one or more of a wired network; a wireless network; a local area network (LAN); and a wide area network (WAN) ;. [0017]

The system may include a resource allocator that allocates shared bandwidth between the subscriber CPE stations. The resource allocator optimizes the quality of service (QoS) of the end user. The resource allocator senses the application. [0018]

The present invention relates to a quality of service (QoS) sensitive, wireless point-to-multipoint telecommunications system. The system communicates wirelessly with a wireless base station connected to the first data network, one or more host workstations connected to the first data network, and the wireless base station via a shared band using a packet-centric protocol. Shared bandwidth between one or more CPE (customer premise equipment) stations and one or more subscriber workstations connected to each subscriber CPE station via a second network. It consists of a resource allocation device to be allocated and a resource allocation device. Resource allocation is performed to optimize end-user QoS. [0019]

The resource allocation means includes an analyzer and a scheduler that analyze and schedule the IP flow through the shared bandwidth, and the IP flow includes the transmission control protocol / Internet Protocol (TCP / IP) flow and the user datagram protocol /. Includes one or more of the Internet Protocol (UDP / IP) flows. [0020]

The analyzer and the scheduler include an identification element that identifies the IP flow, a characteristic element that characterizes the IP flow, and a classification element that classifies the IP flow, and the analyzer and the scheduler include the IP flow. Can be provided with a prioritizing element that prioritizes. [0021] [0021]

The identification element includes an analyzer that analyzes packet header fields and an identification element that identifies a new IP flow and an existing IP flow. The analyzer includes a buffer that buffers packets of the IP flow, a data extraction device that extracts data from the packet header fields of each packet, and a packet header field analyzer that analyzes the packet header fields. .. The data extraction device can include means for determining whether the packet version of the IP flow is IPv.4 or IPv6, and means for analyzing the packet. The packet header field analyzer comprises means for determining the source application type, and means for storing and retrieving the source application for the source address from the source application packet header table; a type of service (TOS) packet. A means of determining the source application from header fields; Differentiated Services (DiffServ) services) means to determine the source application from the packet header field; can consist of one or more of them. [0022]

The characterization element is a means for determining the existence time for determining whether or not the existence time of the packet exceeds the threshold time, and a means for predicting the destruction of the client application IP flow based on the existence time of the packet. , The QoS determination means for determining the QoS requirement of the new IP flow, and the means for determining the subscriber identification of the subscriber CPE station related to the new IP flow can be provided. [0023]

The time-to-live determination means includes means for analyzing an expiration time (TTL: time to live) packet header field to determine the existence time of the packet, and the QoS determination means is a source address; a destination address; UDP. Determine the QoS requirement based on one or more of the port numbers; The QoS determination means can include means for storing and retrieving the QoS requirements of the IP flow from the IP flow QoS requirements table. [0024]

The classification element can include means for associating a packet of the IP flow with an existing IP flow. The classification element can include a classification means for classifying the packets of the new IP flow into QoS class grouping. [0025]

The classification means may include means for determining and considering the QoS class grouping of the IP flow. The classification means may include means to consider the optional Differentiated Services (DiffServ) field priority marking for the IP flow. The classification means may include means to consider the optional type of service (TOS) field priority marking for the IP flow. The prioritizer may include means for considering hierarchical class based priorities (HCBPs) for the IP flow. The prioritizer may include means for considering the virtual private network (VPN) priority for the IP flow. The prioritizer is a service level agreement (SLA) for the IP flow. It is possible to provide a means for considering the priority based on the agreement). The prioritization element may include means for considering the type of service (TOS) priority for the IP flow. The prioritizer may include means for considering differentiated services (DiffServe) for the IP flow. [0026]

The present invention wirelessly communicates with a wireless base station connected to the first data network, one or more host workstations connected to the first data network, and the wireless base station via a shared band by a packet-centric protocol. One or more customer premises equipment (CPE) stations, one or more subscriber workstations connected to each subscriber CPE station via a second network, and the quality of service (QoS) of the end user. A means for analyzing and scheduling a resource allocation device for allocating a shared bandwidth between subscriber CPE stations and a means for analyzing and scheduling an Internet Protocol (IP) flow over the shared radio bandwidth. Relates to packet-centric wireless point-to-multipoint telecommunications systems, including identification means for identifying IP flows. [0027]

The IP flow may be a transmission control protocol / internet protocol (TCP / IP) flow. The IP flow may also be a User Datagram Protocol / Internet Protocol (UDP / IP) flow. The identification means may include means for analyzing one or more packet header fields of the IP flow and means for distinguishing between the new IP flow and the existing IP flow. [0028]

The analytical means can be located at the subscriber CPE station for an uplink radio connection from the subscriber CPE station to the radio base station. The identification means can be located at the subscriber CPE station for an uplink radio connection from the subscriber CPE station to the radio base station. [0029]

The analytical means can be located at the pre-radio base station for a downlink radio connection from the radio base station to the subscriber CPE station. The identification means can be located at the front radio base station for downlink radio connection from the radio base station to the subscriber CPE station. [0030]

The means for analyzing the packet header field may include a means for buffering the packet of the IP flow, an extraction means for extracting data from the packet header field for each packet, and a means for analyzing the packet header field. it can. The extraction means may include means for determining whether the packet version is IPv.4 or IPv6, and means for analyzing the packet header field of the IP flow. [0031]

The analysis means can be provided with a determination means for determining the source application type, and the determination means stores the source application type in the source application packet header table or retrieves the source application type from the source application packet header table. And a means for determining the source application from the type of service (TOS) packet header field. [0032]

The present invention wirelessly communicates with a wireless base station connected to the first data network, one or more host workstations connected to the first data network, and the wireless base station via a shared band by a packet-centric protocol. One or more customer premises equipment (CPE) stations, one or more subscriber workstations connected to each subscriber CPE station via a second network, and the quality of service (QoS) of the end user. A packet-centric radio with a resource allocation device that optimizes and allocates shared bandwidth between subscriber CPE stations and means for analyzing and scheduling Internet Protocol (IP) flows over the shared radio bandwidth. Regarding point-to-multipoint telecommunications systems. The means of analysis include characterization means that characterize the IP flow. [0033]

The characterization means can include an existence time determining means for determining whether or not the existence time of a packet exceeds a threshold time. The existence time determination means is a means for analyzing a time to live (TTL) packet header field to determine the packet, or a means for predicting the destruction of the application IP flow based on the existence time of the packet. Can be prepared. [0034]

The characterization means may include QoS determination means for determining the QoS requirements of the IP flow when the IP flow is a new IPP flow. The characterization means may include means for determining the subscriber CPE identification of the subscriber CPE station associated with the IP flow if the IP flow is a new IPP flow. The QoS determination means may include means for determining the QoS requirement based on one or more of the source address; the destination address; the UDP port number ;. The QoS determination means can include means for storing and retrieving the QoS requirements of the IP flow from the IP flow QoS requirements table. [0035]

The present invention relates to an IP flow classification system used in a wireless telecommunications system. More specifically, it relates to an IP flow classification system that groups IP flows in a packet-centric wireless point-to-multipoint telecommunications system. [0036]

The classification system wirelessly communicates with a radio base station connected to the first data network, one or more host workstations connected to the first data network, and the radio base station via a shared band by a packet-centric protocol. One or more CPE (customer premise equipment) stations, one or more subscriber workstations connected to each subscriber CPE station via a second network, and the quality of service of the end user ( A resource allocation device that optimizes QoS) and allocates shared bandwidth between subscriber CPE stations, and a means of analyzing and scheduling Internet Protocol (IP) flow over the shared radio bandwidth. Includes a , and . The means of analysis include a classification element that classifies IP flows. [0037]

In one embodiment, the classification element comprises means of associating a packet of the IP flow with an existing IP flow. The classification element can include a QoS grouping device that groups packets of new IP flows into QoS class groupings. The QoS grouping device can be provided with means for determining and considering the QoS class grouping of the IP flow. The QoS grouping device may include an optional Differentiated Services (DiffServ: differentiated services) device that considers the optional Differentiated Services (DiffServ) field priority marking for the IP flow. The QoS grouping device may include an optional type of service (TOS) device that considers the optional type of service (TOS) field priority marking for the IP flow. [0038]

The present invention relates to an IP flow prioritization system used in a wireless telecommunications system. The system wirelessly communicates with a radio base station connected to the first data network, one or more host workstations connected to the first data network, and the radio base station via a shared band by a packet-centric protocol. One or more CPE (customer premise equipment) stations, one or more subscriber workstations connected to each subscriber CPE station via a second network, and end-user service quality (QoS). A resource allocation device that optimizes and allocates shared bandwidth between subscriber CPE stations, an IP flow analyzer that analyzes the IP (Internet Protocol) flow connected to the resource allocation device, and an IP that schedules the shared radio bandwidth. The IP flow scheduler includes a flow scheduler, and the IP flow scheduler provides an IP flow prioritization element. [0039]

In one embodiment of the invention, the prioritizing element prioritizes the IP flow based on the HCBP (hierarchical class based priorities) priority of the IP flow. It can have a prioritizing element. In one embodiment of the invention, the HCBP prioritizing element may comprise a class-based priority limit that establishes a limit for each of the HCBP priorities. [0040]

In one embodiment of the invention, the prioritizer is a virtual private network (VPN) prioritizer that prioritizes the IP flow based on a source that is a VPN (virtual private network). Can be provided. In one embodiment of the invention, the virtual private network (VPN) prioritizer can prioritize all of the VPN IP flows. In another embodiment of the invention, the virtual private network (VPN) prioritizer prioritizes one or more of a VPN flow of a particular IP flow type; and a VPN IP flow from a VPN type; be able to. In one embodiment of the invention, the VPN type can include a DEN (directory enabled networking) table management scheme type. [0041]

In one embodiment of the invention, the prioritizer prioritizes the IP flow based on the SLA level of the subscriber source of the IP flow, based on a service level agreement (SLA). It can be equipped with a prescription element. In certain embodiments of the invention, said SLA levels include, for example, premium levels, standard levels, and value levels. In one embodiment of the invention [0042]

The prioritization device can include a type of service (TOS) prioritizing element that prioritizes the IP flow based on the TOS marking of the packet of the IP flow. [0043]

In one embodiment of the invention, the prioritizing element comprises a Differentiated Services (DiffServ) prioritizing element that prioritizes the IP flow based on the Diffserv marking of the packet of the IP flow. Can be done. [0044]

In one embodiment of the invention, the prioritizing element sets a reservation policy limit based on the IP flow priority, ensuring a fair distribution of shared bandwidth (WFP) weighted fair priority. ) Can have a prioritizing element. [0045]

The present invention wirelessly communicates with a wireless base station connected to the first data network, one or more host workstations connected to the first data network, and the wireless base station via a shared band by a packet-centric protocol. One or more customer premises equipment (CPE) stations, one or more subscriber workstations connected to each subscriber CPE station via a second network, and the quality of service (QoS) of the end user. A resource allocation means (manner) that optimizes and allocates shared bandwidth between subscriber CPE stations, and a means (manner) that analyzes and schedules the Internet Protocol (IP) flow over the shared radio bandwidth. The means to be analyzed relates to a packet-centric radio point-to-multipoint telecommunications system, including means for prioritizing IP flows based on the priority of the customer premises equipment (SLA) of the SLA subscriber. [0046]

Alternative functions of the system include means for analyzing SLAs of IP flows (way) and means for prioritizing IP flows based on one or more subscriber-defined parameters (way). SLA levels can include premium service levels, normal service levels, and value service levels. The system uses different SLA levels to (1) traffic rates between SLA subscribers, (2) network availability between SLA subscribers, (3) bandwidths for each of the SLA subscribers, (4). It can differentiate SLA subscriber error rates, (5) SLA subscriber latency guarantees, and (6) SLA subscriber jitter guarantees. [0047]

A wireless base station connected to the first data network, one or more host workstations connected to the first data network, and one or more wireless base stations that communicate wirelessly with the wireless base station via a shared band using a packet-centric protocol. Optimize the quality of service (QoS) of the end user with a CPE (customer premise equipment) station and one or more subscriber workstations connected to each subscriber CPE station via a second network. The analysis includes a resource allocation means (manner) for allocating shared bandwidth between subscriber CPE stations, and a means (manner) for analyzing and scheduling an Internet protocol (IP) flow over the shared radio bandwidth. The means to do so also disclose the scheduling method used for packet-centric wireless point-to-multipoint telecommunications, including the step of prioritizing IP flows based on the priority of the SLA subscriber's service level contract (SLA). To. [0048]

In one embodiment, the method may also include analysis of the flow SLAs. Yet another embodiment includes the step of prioritizing IP flows based on one or more subscriber-defined parameters. The method may also include a step of prioritizing the premium service level, a step of prioritizing the normal service level, and a step of prioritizing the value service level. [0049]

The present invention wirelessly communicates with a radio base station connected to the first data network, one or more host workstations connected to the first data network, and the radio base station via a shared band by a packet-centric protocol. One or more customer premises equipment (CPE) stations, one or more subscriber workstations connected to each subscriber CPE station via a second network, and end-user service quality (QoS). It relates to a service quality (QoS) sensitive, wireless point-to-multipoint communication system consisting of a resource allocation device that optimizes and allocates shared bandwidth between subscriber CPE stations. [0050]

The resource allocation device can include an allocation means for allocating a future slot of a transmission frame to a data packet of a transmission frame transmitted via the radio medium. The allocation means includes a means for applying the pre-reservation algorithm, a first reservation means for reserving the first slot for the first data packet of the Internet Protocol (IP) flow of the future transmission frame based on the pre-reservation algorithm, and the like. A second reservation means for reserving a second slot for a second data packet of the IP flow of the IP flow of a subsequent transmission frame in time to the future transmission frame based on the advance reservation algorithm can be provided. The second data packet is placed in the second slot at the same time that one data packet is placed in the first slot. In one embodiment, there is a periodic variation between the first data packet being placed in the first slot and the second data packet being placed in the second slot. In another embodiment, there is an aperiodic variation between the first data packet being placed in the first slot and the second data packet being placed in the second slot. [0051]

The resource allocation device can be provided with means for indicating priorities (HCBPs: hierarchical class based priorities) based on the hierarchical class regarding the IP flow. The resource allocating device may include means for considering the virtual private network (VPN) priority with respect to the IP flow. The resource allocating device may include means for considering priorities based on a service level agreement (SLA) for the IP flow. The resource allocator may include means for considering the type of service priority for the IP flow. The resource allocation device can be provided with a means for considering differentiated services (DiffServe) regarding the IP flow. [0052]

The present invention provides a wireless telecommunications network with excellent quality of service. The method of allocating the future slot of the transmission frame to the data packet of the transmission frame transmitted through the wireless telecommunications network system is the step of applying the advance reservation algorithm and the Internet protocol of the future transmission frame based on the advance reservation algorithm. The step of reserving the first slot for the first data packet of the IP) flow and the second for the second data packet of the IP flow of the transmission frame temporally following the future transmission frame based on the advance reservation algorithm. A step of reserving a slot is provided, and the second data packet is arranged in the second slot at the same time as the first data packet is arranged in the first slot. [0053]

In one embodiment, there is a periodic variation between the first data packet being placed in the first slot and the second data packet being placed in the second slot. In another embodiment, there is an aperiodic variation between the first data packet being placed in the first slot and the second data packet being placed in the second slot. [0054]

The pre-booking algorithm determines whether the IP flow is jitter sensitive. [0055]

Providing isochronous data packets to an electrical communication system. An electric communication system communicates wirelessly with a wireless base station connected to the first data network, one or more host workstations connected to the first data network, and the wireless base station via a shared band by a packet-centric protocol. One or more CPE (customer premise equipment) stations to perform, one or more subscriber workstations connected to each subscriber CPE station via a second network, and the quality of service (QoS) of the end user. ) Optimizes and allocates shared bandwidth between subscriber CPE stations, and consists of a pre-booking algorithm applied to the Internet Protocol (IP) flow that follows the IP flow based on the algorithm. Reserve subsequent slots for one or more future transmit frames in an isochronous manner. There may or may not be periodic fluctuations during the subsequent reservation of the subsequent slot, and the algorithm determines whether the IP flow is jitter sensitive. [0056]

A system that allocates a future slot of a transmission frame to a data packet of a transmission frame transmitted via a wireless medium is a means of applying a pre-reserved algorithm and a first of the Internet Protocol (IP) flows of the future transmission frame based on the algorithm. A means for reserving the first slot for one data packet and a means for reserving the second slot for the second data packet of the IP flow of the transmission frame following the future transmission frame in time based on the algorithm. The second data packet is placed in the second slot at an isochronous time when the first data packet is placed in the first slot. The arrangement may or may not be periodic. The pre-booking algorithm is used to determine if the IP flow is jitter sensitive. [0057]

The system provides isochronous data packets to the telecommunications system. An electric communication system communicates wirelessly with a wireless base station connected to the first data network, one or more host workstations connected to the first data network, and the wireless base station via a shared band by a packet-centric protocol. One or more CPE (customer premise equipment) stations to perform, one or more subscriber workstations connected to each subscriber CPE station via a second network, and the quality of service (QoS) of the end user. ) Is optimized and the resource allocation means for allocating the shared bandwidth between the subscriber CPE stations, and the system is composed of the means for applying the pre-booking algorithm to the Internet Protocol (IP) flow and the IP flow based on the above algorithm. Includes means for isochronically reserving subsequent slots in one or more future transmission frames that follow. There may or may not be periodic fluctuations during the subsequent reservation of the subsequent slot, and the algorithm determines whether the IP flow is jitter sensitive. [0058] [0058]

The present invention provides a wireless telecommunications network with excellent quality of service. A system that translates Internet-prioritized Internet Protocol (IP) flows into radio bandwidth resource allocation is a radio base station connected to the first data network and one or more host work connected to the first data network. A station, one or more CPE stations that wirelessly communicate with a radio base station over a shared band using a packet-centric protocol, and one or more subscribers connected to each subscriber CPE station via a second network. Includes a wireless point-to-multipoint (PtMP) transmission system consisting of a workstation and a resource allocator that allocates shared bandwidth between subscriber CPE stations to identify the IP flow identifier in the IP priority packet header. Includes a TCP / IP, UDP / IP flow (IP flow) analyzer that classifies IP flows, and an IP flow scheduler that prioritizes IP flows in consideration of IP priority header identification information. [0059]

The IP Priority Packet Header IP Flow Identification Information includes means for determining and considering QoS class grouping for IP flows. The IP Priority Packet Header IP Flow Identification also considers the Type of Service (TOS) field priority marking. The type of service (TOS) field priority marking is compatible with the Internet Engineering Task Force (IETF) RFC 1992b, and IETF RFC 1349. Markings include minimum delay marking, maximum throughput marking, maximum reliability marking, minimum monetary cost marking, and standard service marking. In addition, the system may process IP flow identification information and consider optional Differentiation Service (DivffServ) field priority markings, such as those compatible with Internet Engineering Task Force (IETF) RFC 2474 and IETF RFC 2475. it can. [0060]

In one embodiment, the IP priority packet header IP flow identifier provides means to consider Resource Reservation Protocol (RSVP) messages and objects. The RSVP message is a path message; [0061]

Reservation (Resv); path teardown message; resv teardown message; path error message; and confirmation message; can be included. The other RSVP protocol objects are null; session; RSVP_hop; time_values; style; flowspec; sender_template. [0062]

It can include sender_Tspec; Adspec; Error_Spec; Policy_data; Integrity; Scope; and Resv_Confirm ;. The RSVP markings are compatible with the Internet Engineering Task Force (IETF) RFC 2205. [0063]

The present invention wirelessly communicates with a wireless base station connected to the first data network, one or more host workstations connected to the first data network, and the wireless base station via a shared band by a packet-centric protocol. Between one or more CPE stations, one or more subscriber workstations connected to each subscriber CPE station via a second network, and end user service quality (QoS) optimized. The present invention relates to a packet-centric radio point-to-multipoint electric communication system including a resource allocation means for allocating a shared bandwidth to a network and a means for analyzing and scheduling an Internet protocol (IP) flow via the shared radio bandwidth. [0064]

Scheduling means include prioritization means that prioritize IP flows based on virtual private network (VPN) priorities. The system can be provided with a means of analyzing the virtual private network (VPN) priority of IP flows, or all VPN IP flows. The system can be provided with means for prioritizing IP flows based on one or more subscriber-defined parameters. In the system, the VPN can include a DEN (directory enabled networking) table management scheme. VPNs can be run using the Point-to-Point Tunneling Protocol (PPTP). [0065]

Further included is a method of performing the above. [0066]

The present invention relates to a packet-centric radio point-to-point (PtP) telecommunications system. The system communicates wirelessly with a wireless base station connected to the first data network, one or more host workstations connected to the first data network, and the wireless base station via a shared band using a packet-centric protocol. It consists of one or more CPE (customer premise equipment) stations and one or more subscriber workstations connected to each subscriber CPE station via a second network. The packet-centric protocol may be a transmission control protocol / Internet protocol (TCP / IP). The packet-centric protocol may be a user diagram protocol / internet protocol (UDP / IP). [0067]

The system may include resource allocating means for allocating shared bandwidth between said subscriber CPE stations. The resource allocation means can optimize the quality of service (QoS) of the final user. [0068]

The wireless communication medium can be composed of one or more of a radio frequency (RF) communication medium; a cable communication medium; and a satellite communication medium ;. The telecommunications media also include time division multiple access (TDMA) access methods; time division multiple access / time division duplex (TDMA / TDD) access methods; code division multiple access (CDMA) access methods; and frequency division multiple access. It can consist of telecommunications access methods, including one or more of (FDMA) access methods; [0069]

The first data network can consist of one or more of a wired network; a wireless network; a local area network (LAN); and a wide area network (WAN) ;. [0070]

The system can be equipped with a resource allocator that allocates shared bandwidth between the subscriber CPE stations. The resource allocator optimizes the quality of service (QoS) of the end user. The resource allocator can sense the application. [0071]

The present invention comprises a cable radio base station that communicates with a first data network using a packet-centric protocol, one or more host workstations that communicate with the first data network using the packet-centric protocol, and a coaxial cable communication medium. One or more subscriber CPE (customer premise equipment) stations connected to the radio base station via a shared bandwidth via the packet-centric protocol and each subscriber CPE via a second network. It relates to a packet-centric broadband coaxial cable point telecommunications system consisting of one or more subscriber workstations connected to a station by a packet-centric protocol. The packet-centric protocol may be a transmission control protocol / Internet protocol (TCP / IP). The packet-centric protocol may be a user diagram protocol / internet protocol (UDP / IP). [0072]

The system may include a cable resource allocator that allocates shared bandwidth between said subscriber CPE stations. The resource allocator can optimize the quality of service (QoS) of the end user. The coaxial cable communication medium can include radio frequency data communication over the coaxial cable, and one or more cable modems modulate and demodulate signals transmitted through the medium. The cable modem can be DOC / SYS compliant. The QoS optimized cable resource allocator system is an IP flow identification element; an IP flow characterization element; [0073]

It can include IP flow classification elements; and IP flow prioritization elements ;. [0074]

The coaxial cable communication medium includes a time division multiple access (TDMA) access method; a time division multiple access / time division duplex (TDMA / TDD) access method; a code division multiple access (CDMA) access method; and a frequency division multiple access. It can consist of telecommunications access methods, including one or more of (FDMA) access methods; The first data network may be a wired network; a wireless network; a local area network (LAN); and a wide area network (WAN) ;. The second network can consist of one or more of a wired network; a wireless network; a local area network (LAN); and a wide area network (WAN) ;. [0075]

The resource allocator can be application aware. The system can be a point-to-point (PtP) network. [0076]

The present invention relates to a method of allocating shared radio bandwidth to a packet-centric radio point-to-multipoint telecommunications system. The method can include allocating the shared bandwidth between a radio base station and one or more customer premises equipment (CPE) stations. The method can include the step of dynamically allocating the shared bandwidth. The method is the step of (1) allocating the frame of the shared bandwidth in the uplink direction from the subscriber CPE station to the radio base station, or (2) the said of the shared bandwidth. A step of allocating a shared bandwidth for each frame can be included, such as by allocating frames from the radio base station to the subscriber CPE station in the downlink direction. [0077]

The method can further include allocating the shared bandwidth for each subframe within the frame, i.e., the uplink direction of the subframe of the shared bandwidth from the subscriber CPE station to the radio base station. A step of allocating the subframe of the shared bandwidth to the subscriber CPE station in the downlink direction from the radio base station can be included. [0078]

In one embodiment, the method can include allocating the shared bandwidth for each slot in the frame, i.e., allocating the slot for the shared bandwidth from the subscriber CPE station to the radio base station. The step of allocating the slot of the shared bandwidth in the uplink direction, and the step of allocating the slot of the shared bandwidth in the downlink direction from the radio base station to the subscriber CPE station can be included. [0079]

In one embodiment, the step of allocating the shared bandwidth for each subslot in the frame can be included, i.e., the subslot of the shared bandwidth is upgraded from the subscriber CPE station to the radio base station. A step of allocating the subslot of the shared bandwidth in the link direction and a step of allocating the subslot of the shared bandwidth in the downlink direction from the radio base station to the subscriber CPE station can be included. [0080] [0080]

The method can include allocating the shared bandwidth to one or more control packets. The method is: a step of allocating a downstream affirmative response slot; a step of allocating a reserved request slot; a step of allocating an operations data slot; a step of allocating an upstream affirmative response slot; a step of allocating an affirmative response request slot; a frame description. It can include steps to assign children; and; to assign commands and control slots. [0081]

The method can include allocating shared bandwidth to one or more data packets. The method can include the step of allocating the shared bandwidth in the uplink direction; and the step of allocating the shared bandwidth in the downlink direction; [0082]

The present invention relates to a method of providing quality of service (QoS) sensitive, wireless point-to-multipoint telecommunications to a telecommunications system. Telecommunications communicates wirelessly with a wireless base station connected to a primary data network, one or more host workstations connected to a primary data network, and a wireless base station over a shared band using a packet-centric protocol. It consists of one or more CPE stations to perform and one or more subscriber workstations connected to each subscriber CPE station via a second network. [0083]

The method of the present invention can include the step of allocating the shared bandwidth between the subscriber CPE stations in a way that optimizes the QoS of the end user. [0084]

The method includes the steps of analyzing and scheduling IP flows over the shared radio bandwidth. This is the step of identifying the IP flow and [0085]

It includes a step of characterizing the IP flow, a step of classifying the IP flow, and a step of prioritizing the IP flow. [0086]

The identification step can include a step of analyzing the packet header field and a step of identifying a new IP flow and an existing IP flow. This includes a step of buffering the packet of the IP flow, a step of extracting data from the packet header field of each packet, and a step of analyzing the packet header field. It also includes a step of determining whether the packet version of the IP flow is IPv.4 or IPv6, and means for analyzing the packet. It also includes the step of determining the source application type, saving and retrieving the source application for the source address from the source application packet header table; sending from the type of service (TOS) packet header field. Steps to determine the original application; Differentiated Services (DiffServ: differentiated) services) Consists of one or more of the steps to determine the source application from the packet header field; The identification step can also include storing an existing IP flow in the IP flow identification data table or retrieving it from the IP flow identification data table. [0087]

The method also includes a step of determining whether the packet existence time exceeds the threshold time, a step of predicting the destruction of the client application IP flow based on the packet existence time, and the new IP. It can include a step of determining the QoS requirements of the flow and a step of determining the subscriber identification of the subscriber CPE station associated with the new IP flow. It analyzes the time to live (TTL) packet header field to determine the time to live of the packet. [0088]

The method can also include a step of determining the QoS requirements of the new IP flow based on one or more of the source address; the destination address; the UDP port number ;. The method can also include a step of classifying the packets of the new IP flow into QoS class groupings. This is the step of determining and considering the QoS class grouping of the IP flow, the step of considering the optional Differentiated Services (DiffServ) field priority marking for the IP flow, and the optional service type for the IP flow. (TOS: type of service) A step that considers field priority marking, a step that considers hierarchical class based priorities (HCBPs) for the IP flow, and a virtual private network (VPN) for the IP flow. : virtual private Network) Priority Consideration Step, Priority Consideration Based on Service Level Agreement (SLA) for IP Flow, Service Level Agreement (TOS: Type of Service) Priority Consideration for IP Flow Steps to consider, consider differentiated services (DiffServe) for the IP flow, prioritize the IP flow based on the service level agreement (SLA) priority of the SLA subscriber, A step of analyzing the SLA of the IP flow can be included. [0089]

The present invention relates to a method of incorporating a Differentiated Services (DiffServe) marked IP flow into a quality of service (QoS) priority in a point-to-multipoint (PtMP) transmission system. The system wirelessly communicates with a radio base station connected to the first data network, one or more host workstations connected to the first data network, and the radio base station via a shared band by a packet-centric protocol. Consists of one or more CPE stations, one or more subscriber workstations connected to each subscriber CPE station via a second network, and a resource allocator that allocates shared bandwidth between subscriber CPE stations. Will be done. The method of the present invention includes a step of analyzing an IP flow for Differentiated Services (DiffServ) marking and a step of scheduling the IP flow in consideration of the Differentiated Services (DiffServ marking). [0090]

The IP flow can include one or more of Transmission Control Protocol / Internet Protocol (TCP / IP) and User Diagram Protocol / Internet Protocol (UDP / IP). The analysis step can include identifying IP flows with DiffServ marking, characterizing IP flows with DiffServ marking, and classifying IP flows with DiffServ marking. The scheduling step may include prioritizing the IP flow taking into account DiffServ marking and IP priority header identification information. The prioritizing step can include weighted fair priority (WFP). [0091]

The prioritization steps are prioritized based on the priority given by the IP flow hierarchy class; prioritized based on the service level contract (SLA) class priority; virtual private network (VPN). You can have one or more of the following steps: network) prioritizing based on subscribers; virtual private network (VPN) prioritizing based on subscriber class priorities; .. [0092]

The identification step can include one or more steps of analyzing one or more packet header fields of the IP flow; identifying a new IP flow from an existing IP flow. The step of analyzing the packet header field is a step of buffering the packet of the IP flow; a step of extracting identification information from the packet header field of each packet; a step of analyzing the identification information from the packet header field; It is possible to have one or more steps in. The extraction step can include a step of determining whether the packet version is IPv.4 or IPv6, and a step of analyzing the packet header field of the IP flow. The analysis step can include a step of determining the source application type. The analysis step is a Differentiated Services (DiffServ) services) Can be equipped with steps to consider field priority marking. The Differentiated Services Field Priority Marking can be compatible with the Internet Engineering Task Force (IETF) RFC 2474. The Differentiated Services Field Priority Marking can be compatible with the Internet Engineering Task Force (IETF) RFC 2475. [0093]

The classification step can include associating a packet of the IP flow with an existing IP flow. The classification step can include a grouping step that groups the packets of the new IP flow into QoS priority classes. The grouping step may include a step that considers the Differentiated Services marking for the IP flow. The prioritization step may include a step that considers DiffServ marking for the IP flow. [0094]

(Embodiment of the Invention) VII. Environmental example [0095]

The present invention will be described with reference to an environmental example. An example of this environment uses a fixed wireless point-to-multipoint (PtMP) connection to transmit packetized data information, such as IP phones, videos, and data received from a telecommunications carrier. Telecommunications carriers mentioned in this document are not only those in the United States such as ILEC, CLEC, IXC, NGT, and ESP (Extended Service Providers) (see "Definition" below), but also PTTs known to those skilled in the art. Can include international operators such as, NE. In addition, the telecommunications systems referred to herein can include not only domestic systems used by operators such as ILEC, CLEC, IXC, ESP (Extended Service Providers), but also global systems known to those of skill in the art. [0096]

In a preferred embodiment, the traffic comes from a wide area network (WAN) connection. Data traffic is received from the data network through the network router and can be demolished from Internet Protocol (IP) format, for example to Point-to-Point Protocol (PPP). Network routers are, for example, general-purpose computers such as SUN workstations running routing software, or CISCO (San Jose, CA), ASCEND (Alameda, CA), NETOPIA (Alameda, CA), 3COM (Santa, Santa, It can include dedicated routing devices for various models such as Clara). [0097]

In an alternative embodiment, a virtual private networking protocol such as the Point-to-Point Tunneling Protocol (PPTP) can be used to create a "tunnel" between a remote user and a corporate data network. Tunnels allow network administrators to extend virtual private networks (eg the Internet) from servers (eg Windows NT® servers) to data networks. [0098]

The present invention will be described herein using an example of the environment, but the description in this document is for the purpose of explaining the present invention only, and does not limit the present invention. In fact, readers who read the instructions in this book will find out how to implement the invention in an alternative environment. VIII. Definition [0099]

Table 1 below defines common telecommunications terms. The defined terms will be used herein in the description of the present invention.

[table 1]

<img file="JP2003521138A_D0001.tif" /><img file="JP2003521138A_D0002.tif" /><img file="JP2003521138A_D0003.tif" /><img file="JP2003521138A_D0004.tif" /><img file="JP2003521138A_D0005.tif" /><img file="JP2003521138A_D0006.tif" />IX. Introduction A. Quality of Service (QoS) in a wireless environment [0100]

The concept of quality of service (QoS) is one of the most difficult and incomprehensible topics in data networking. Although commonly used in data networking, QoS has many different uses and definitions and is confusing in the strict, or quantitative sense. Attempts to measure or identify QoS numerically so that device or network performance can be compared creates further confusion. [0101]

The general data networking QoS confusion is similar and even greater for wireless data communications. Wireless transmission has a higher unique BER (Bit Error Rate) than wired transmission. When adding a PtMP (Point to Multipoint) topology in which multiple users share a wireless medium, it is preferable to define QoS so that multiple complex factors of wireless data communication can be clearly handled. [0102]

In order to clearly define the QoS applied to wireless data communication, it is useful to know the nature of the problem that QoS needs to solve. Some of the problems of wireless data communication are shared with wired data communication, but many are specific to wireless data communication and are more pronounced than the problem of wired data communication. In wireless broadband access systems, the problem of high quality transmission is slightly more complicated than in wired analog. Problems encountered by wireless data transmission, such as in wired communication, include, for example, slow access to peripherals, data errors, "dropouts", unwanted retransmissions, traffic congestion, out-of-order data packets, and waits. There is time, and jitter. In addition to these issues, wireless transmission includes, for example, inherently high BER (bit error rate), bandwidth limitation, user contention, wireless failure, and TCP traffic rate management. QoS-sensitive wireless systems preferably address all of these issues. [0103]

Data network users or subscribers encounter difficulties in several ways. One of the difficulties of networks is the lack of availability of networks. Depending on the access technology used, this may include a "modem not responding" condition, a "network busy" condition, or an unexpected sudden "disconnection" of a network connection. In such a state, it cannot be said that QoS is high. Once network connectivity is established, slow traffic due to congestion, local access bottlenecks, and network failures can be slow web page loading, slow file transfers, or audio / video audio / video streaming for multimedia applications. You can experience it due to deterioration of quality. Degradation of streaming quality in multimedia applications can be caused by high "jitter" or large and rapid fluctuations in latency, which can cause sessions to be interrupted, distorted, or terminated. Data errors can occur due to various conditions, which can be catastrophic in cases such as spreadsheet transfers. It is desirable to minimize or eliminate such data communication network problems. 1. Quality [0104]

In a data network, quality usually refers to the process of transmitting data reliably and in a timely manner. Reliable and timely depends on the nature of the traffic being handled. These terms may also relate to data loss limits, data accuracy prospects, data latency fluctuation limits (called jitter), data retransmission limits, and data packet order reversal limits. Therefore, since QoS is a complex concept, a complex mechanism for implementing it may also be required. [0105]

QoS is a relative term and has different meanings depending on the user. Temporary users who occasionally browse the web but do not have FTP (File Transport Protocol) file downloads or real-time multimedia sessions, and multiple FTP file downloads of large databases or financial files for H.323 video conferencing and The definition of QoS is different from that of power users who frequently use IP phones. Users can also pay premium rates (so-called Service Level Agreements) with high network usage, low latency, and low jitter, while other users occasionally web on weekends. You can also pay a low fee just for surfing. For this reason, it may be best to understand the network performance that is most important to a particular user, and the continuous QoS defined by the user's SLA. Maximizing the convenience of the end user is a fundamental requirement for providing wireless QoS. 2. Service [0106]

In data networking, a service is defined as the type of connection from one end of the network to the other. Previously, it was also defined by protocol, such as IBM's SNA (System Network Architecture), Novell's IPX, and Digital's DECnet. However, TCP./IP (ie User Datagram Protocol (UDP)) has been selected as the protocol, and the number of users using it is overwhelmingly large. This trend is expected to continue in the future. This allows services to be defined for a particular TCP / IP connection or outbound type. Types of this service can include, for example, FTP file transfer, email traffic, HTTP (Hypertext Transport Protocol) traffic, H.323 video conferencing sessions, and so on. It is desirable for the QoS mechanism to handle these various types of services, as well as the various quality types mentioned above. 3. QoS as a mechanism [0107]

QoS can be thought of as a mechanism for selectively allocating scarce networking, transmit, and communication resources to various different classes of network traffic with appropriate priority levels. Modifying behavior according to the nature of data traffic, user requirements, network conditions, traffic source and destination characteristics, all of which are instant (instant) given a QoS mechanism. Is ideal. However, ultimately, no matter how the user defines it, it is preferable that the QoS mechanism operates so as to provide the optimum service to the user. Circuit-switched QoS [0108]

In traditional networks generated with a focus on telco voice traffic, data transmission was based on line-centric-defined QoS. By this definition, QoS is asynchronous (ie, data transmission with start and end sequences without a common clock), isochronous (ie, consistent and timely to time-sensitive audio and video network bandwidth). Access) Suggested the ability to carry traffic. Circuit-switched QoS is provided by assigning end-to-end lines exclusively for each connection or service for voice (see Figure 1A) or data. The line-centric QoS mechanism simply provided this line for exclusive use by the user. Of course, this approach takes the line, the line, all transmission channels associated with the line, and the transport medium for the entire duration of the session, regardless of whether data is actually being transmitted for each instant of the session. Allocate itself to just one user. It was generally believed that true QoS could only be achieved by this method. Therefore, designs for wireless broadband access systems have traditionally dedicated wireless radio channels to each particular data connection, regardless of application and whether or not data was actually transmitted within a certain period of time. I have used this approach (see Figure 2A). This line-centric approach to QoS has been enormous in view of the cost of the equipment and the utilization rate of the transmission medium itself. b. Asynchronous Transfer Mode (ATM) QoS [0109]

Telephone companies use traditional voice circuit mechanisms and ATM networking to provide PVC (permanent virtual connection) (that is, VPC (virtual path connection) or VCC (virtual channel connection) provided to unspecified users). And SVC (Switched Virtual Connection) (that is, the logical connection between endpoints established by the ATM network in response to requests based on signal messages received from the last user or other network). , Line-centric QoS can be continued. However, in order to handle new traffic classified as VBR (variable bit rate), CBR (constant bit rate), UBR (unspecified bit rate), etc. New concepts such as mission policy and traffic shaping, and mechanisms such as leaky buckets are needed. [0110]

A virtual circuit is established for each data transmission session, regardless of the purpose of the data or whether the data is transmitted over a predetermined period of time. ATMs provide QoS for broadband network traffic, but the basic design premise of ATMs is that they include wired networks with low BER characteristics, but not wireless media with high BER characteristics. It is not possible to provide true QoS without understanding the characteristics of the traffic carried by the ATM mechanism and the high BER characteristics inherent in the radio. The ATM QoS mechanism does not address the specific challenges associated with wireless communications. c. Packet-switched QoS [0111]

Packet switching is a revolutionary data communication that renews traditional circuit-switched and ATM networking concepts and traditional QoS mechanisms. With packet-switched data communication, it is no longer possible to dedicate a line to a particular data communication session. In fact, the strength of packet-switched lines is the flexibility of the route and the parallel relationship with a given physical network. Therefore, the QoS mechanism does not work as well as the traditional line-centric QoS mechanism. [0112]

It is self-evident that simply providing "sufficient" bandwidth does not provide a sufficient QoS mechanism in packet-switched networks, nor does it provide a sufficient QoS mechanism in wireless broadband access systems. IP flows include "bandwidth-sensitive" IP flows, but also latency-sensitive and / or jitter-sensitive IP flows. Simply providing excess bandwidth, even if cost-prohibited, cannot guarantee real-time or multimedia flow, and timely operation of the application. For IP-centric wireless broadband access systems, the QoS mechanism should be aware of the detailed requirements for each flow of traffic and allocate the necessary system and media resources to carry those flows in an optimal way. d. Summary-QoS mechanism [0113]

Ultimately, the experience of the end user ultimately evaluates QoS. Allocate and regulate systems and media to make IP-centric wireless broadband access systems most accessible to end users. For applications that download the first screen of a web page, data transmission speed is the best measure of QoS. Also, in the case of applications such as spreadsheet downloads and uploads, the minimum transmission error is the best guideline for QoS. In some applications, optimizing both speed and error is the best measure of QoS. In some applications, transmitting packets in a timely manner is the best guideline for QoS. Here, it is important to note that high-speed data transmission does not necessarily mean transmitting packets in a timely manner. For example, if you send an already "too old" data packet immediately, the user may no longer need it because the data packet is already too old. The use of the data itself and the nature of convenience desired by the end user can provide the most reliable criteria for the QoS mechanism. It is desirable for an IP-centric wireless broadband access system to dynamically optimize system operation for each individual IP flow and provide a QoS mechanism that responds appropriately to changes in network load, congestion, and error rates. .. 4. Service Assurance and Service Level Agreements (SLAs) [0114]

Service can be guaranteed. That is, an SLA (Service Level Agreement) that specifies the network availability level and includes access charges based on the specified level can be agreed between the telecommunications service provider and the subscriber. Unfortunately, it is difficult to quantify the degree of network availability for a given time, so it is a rough guide to service performance. It is desirable to use data transmission speed, error rate, retransmission, latency, and jitter as indicators of network availability, but measuring their quantity in real time is a common NSP (Network Service Provider). It is beyond the ability. [0115]

Other service level indicators that network service providers desire are service level agreements (SLAs) that specify guaranteed individual traffic speeds, network availability, bandwidth, error rates, latency, and jitter. There is a need for an IP-centric wireless broadband access system that can provide SLAs that can provide various services and diversify profit margins for service providers. 5. Service class and quality of service [0116]

In order to implement a practical QoS mechanism, a system that can identify the type of traffic or type of service and allocate different levels of system resources to each type is preferred. Traditionally, "class of service" has been referred to as a means of grouping the types of traffic that are allocated and processed with equivalent system and media resources. [0117]

Currently, there are several methods available for wireless network devices to identify a class of service. Examples of methods include traffic shaping, admission control, IP precedence, and differential congestion management. IP-centric wireless broadband access systems use all of these methods to differentiate traffic by service class and map those service classes to a QoS matrix to simplify the operation and management of QoS mechanisms. B. QoS and IP-centric wireless environment [0118]

Point-to-multipoint (PtMP) [0119]

In a point-to-multipoint (PtMP) wireless system similar to the present invention, it is desirable that the QoS mechanism considers not only wired networking but also the wireless environment. As mentioned above, it is desirable to handle radio-specific BERs. If the BER is high, it can be requested that error detection, error correction, and retransmission be performed efficiently. The BER processing mechanism also preferably operates efficiently with a TCP / IP retransmission algorithm so as not to unnecessarily reduce bandwidth utilization. Another radio challenge is competition between users over finite radio bandwidth. Since the system processes service requests from multiple users on a wireless medium that is susceptible to interference and noise, it is difficult to allocate radio bandwidth efficiently. [0120]

As mentioned above, changing from circuit-switched and ATM data networks to packet-switched data networks also affects the definition of QoS mechanisms. The present invention provides a novel QoS mechanism in a point-to-multipoint IP-centric wireless system for packet-switched network traffic. In order for the system to achieve optimal QoS performance, it is preferable to introduce a new approach to the QoS mechanism. Using QoS as a basic guideline for system architecture and design is an existing wireless broadband access system designed with a traditional line-centric or ATM cell line approach, such as that used by Teligent and Winstar. On the other hand, it is an important, substantial and advantageous difference in the IP-centric wireless broadband access system of the present invention. C. IP-centric wireless broadband access QoS and queuing 1. Queue management [0121]

Queuing is generally accepted as a necessary tool for manipulating data communication flows. It is desirable to queue data packets when inspecting or modifying packet headers, determining routes, or outputting data flows to the appropriate ports. However, queuing inevitably causes a delay in the stream of traffic by definition. Delays can be detrimental and can have consequences that are completely contrary to the purpose of queuing. Excessive queuing can delay time-sensitive packets beyond the available time frames, increase RTT (Round Trip Time), resulting in unacceptable jitter, and in some cases. Can also have a negative impact on traffic, such as timing out the data transport mechanism. Therefore, it is advisable to use queuing systematically and refrain from heavy use so that delay-sensitive traffic, such as real-time sessions, does not delay more than necessary. [0122]

In a wireless environment that uses technologies such as TDMA (Time Division Multiple Access) and FEC (Forward Error Detection), it is desirable to use queuing only to enable packet and wireless frame processing. However, when queuing in real-time flow, it is preferable to keep the total delay of the entire real-time traffic below about 20 ms. [0123]

Using queue management as the primary QoS mechanism to provide different services depending on the QoS is an easy and direct way for wireless broadband systems. However, wireless systems are usually more sensitive to delays than wired ones due to bandwidth limitations. Therefore, when providing different services according to QoS, it is necessary to provide a mechanism that goes beyond simple queuing. However, there is still some queuing required, so various queuing methods are still being considered. 2. FIFO (First In First Out) Quing [0124]

FIFO (First In First Out) queuing can be used to buffer data packets in wireless systems as well as in wired systems when the downstream data channel is temporarily congested. In the event of temporary congestion due to sudden traffic, a FIFO queue of appropriate depth can be used to lubricate the data flow to the congested communication segment. However, if the degree of congestion is high, or the period of congestion is relatively long, the FIFO queue reaches its capacity limit, and the network can no longer store packets, the FIFO will drop the packet. become. Discarding a packet is called "packet-tossing". This not only has a detrimental effect on network QoS. Discarding packets causes traffic flow problems and exacerbates the situation because the TCP protocol retransmits the lost packets in a predetermined order. The problem of packet discarding can be minimized by increasing the size of the FIFO buffer and increasing the time it takes to discard the packet. Unfortunately, FIFOs cannot be large enough that packets are old enough that they are no longer needed, or that RTT (response time) is long enough to lose virtual data connections. [0125]

In a wireless broadband environment, FIFO queuing requirements partially depend on the type of RF access method used. In TDMA / TDD (Time Division Multiple Access / Time Division Duplex), it is preferable to queue the data even if it collects enough data to build the transmission data frame. FDMA (Frequency Division Multiple Access) and CDMA (Code Division Multiple Access) are not as "continuous" as TDMA, so there are fewer requirements for FIFO queuing. However, in all wireless access technologies, noise and interference generally contribute to retransmissions, further delays, and adversely affect QoS. [0126]

When using FIFO queuing, a shared wireless broadband system can delay all traffic uniformly. This is the "fairest" method, but it is not always the best method when the goal is to provide high QoS to users. By using different types of queue management, the overall QoS base can be significantly improved. 3. Priority queuing [0127]

The shared wireless broadband environment includes compressed bandwidth segments because the data is transmitted over the RF medium. Therefore, no matter what access technology is used, these systems need to perform some queuing. However, FIFO queuing causes a constant delay for all traffic, regardless of traffic priority or type. Most data environments have a mix of real-time, interactive data, files, data downloads, web page access, and other traffic. Among them are the types of traffic that are most sensitive to delay and jitter. Priority queuing simply sorts data packets by priority and type, allowing latency and jitter-sensitive traffic to move to the top of the queue. [0128]

Unfortunately, a "buffer exhausted" condition occurs when downlink data channel congestion or congestion due to excessive high-priority traffic occurs. Higher priority packets consume most of the buffer capacity, leaving less buffer space for lower priority packets. After all, system resources are devoted to high-priority packets, which results in significant delays for low-priority packets. Some low-priority packets are not only packets that are kept in the buffer for a long time, but also packets that do not reach the buffer. The data flow of those packets is significantly delayed, and applications that use those low-priority packets may be disrupted or even shut down. Due to the nature of such a queuing approach, packets with overall low priority have unpredictable latency and jitter, which adversely affects QoS. [0129]

If the queue size is small, sorting the data in the queue has little effect on improving QoS. In fact, the process of inspecting packet headers to get the information needed to sort the queue can cause significant delays for the data stream. Therefore, for wireless broadband data environments, priority queuing is not much different from FIFO queuing as a QoS mechanism. 4. Class Based Queuing [0130]

You can avoid buffer exhaustion by allocating queue space and system resources to packets according to the packet class. Each class can be assigned the lowest level of service so that high priority data flows do not monopolize all system resources. With this classification approach, there is no data flow that is completely blocked so that the source application can receive information about the traffic rate and TCP-mediated transmission rate adjustment is a smooth traffic flow. To support. [0131]

Although this approach is superior to FIFO queuing in wireless broadband systems, it still cannot solve the negative impact of large numbers of high-priority flows on latency and jitter-sensitive flows. 5. Weighted Fair Queuing [0132]

The fair queuing method provides guaranteed queuing resources for small volume flows, ensuring that the remaining flows are allocated the same amount of resources regardless of volume or priority. This avoids buffer depletion and improves latency and jitter performance somewhat, but abrupt changes in available RF downlink channel bandwidth can make stable performance difficult. is there. [0133]

Providing high-quality services requires more sophisticated QoS mechanisms than simple queue management. D. IP-centric wireless broadband access QoS and TCP / IP 1. TCP / IP [0134]

The TCP / IP protocol stack has become the standard method for transmitting data over the Internet and is gradually becoming the standard for VPNs (Virtual Private Networks). The TCP / IP protocol stack includes not only IP (Internet Protocol), but also TCP (Transmission Control Protocol), UDP (User Datagram Protocol), and ICMP (Internet Control Message Protocol). Given that the TCP / IP protocol stack is the standard network protocol for data communication, the set of optimal QoS for wireless broadband data communication The creation of mechanism) is easier to manage. It is possible to create a QoS mechanism that covers the entire network, including both the wired and wireless parts of the network. By integrating these mechanisms smoothly and transparently with the TCP rate control mechanism, it is possible to provide an end-to-end QoS mechanism that adapts to both the wired and wireless parts of the network. Of course, segments of wired networks that are congested or have other transport problems cannot be resolved by wireless QoS mechanisms. However, the wireless QoS mechanism can optimize the data flow to improve the convenience of the end user if there is no congestion or bottleneck in the wired network. 2. Classification by class [0135]

As mentioned above, data traffic can be processed according to the class of service. When classifying traffic according to its living, data traffic (or a sequence of data packets related to a particular application, function or purpose) can be classified into several classes of service. Classification can be performed according to the identification information contained in the packet header. One method is to analyze, for example, some items contained in the IP packet header that uniquely identify and correlate packets and other packets with packet flows of a particular application, function or purpose. .. At a minimum, the source IP address, source TCP or UDP port, destination IP address, and destination IP or UDP port can associate the packet with a common flow and are used to classify the packet into a class of service. It is possible. [0136]

In order to create a manageable, final number of discrete classes of services, the IP flow is processed by the QoS mechanism in an integrated manner with the given QoS parameters. These classes can be defined to provide common and convenient characteristics in an optimal environment where wireless and wired network segments are combined. 3. Classification by flow [0137]

The final discrete class of service is not computer-intensive, uses relatively little memory and state machines, and is therefore more scalable than the QoS mechanism (or set of parameters) per individual IP flow. Enables the mechanism. However, in network access devices such as point-to-multipoint (PtMP) wireless broadband access systems, the total number of simultaneous IP flows does not exceed 1000. Therefore, since the required amount of overhead can be processed, it is possible to perform QoS classification by flow without relying on the service class. However, a service class that integrates IP flows has advantages in sales, billing, and management. [0138]

Prior to the present invention, flow classification was not used in wireless environments (including radio frequencies transmitted over coaxial cable and satellite communications). 4. Use of IP Precedence for class of service [0139]

IP TOS in the service type field theoretically classifies IP flows into classes of service, as described in the Internet Engineering Task Force (IETF) 1992b. It can be used as a means. IETF RFC1349 proposes a set of 4-bit definitions that have five meanings: minimize latency, maximize throughput, maximize reliability, minimize monetary cost, and standard services. [0140]

These definitions could improve QoS by classifying the types of flows and allocating resources appropriately, which could have significant implications for networks, routers and access devices. However, this proposal was not widely accepted. Some IETF proposals have used this field according to RSVP (Resource Reservation Protocol) to improve network packet processing. [0141]

TOS (Service Type) has been part of the TCP / IP specification for many years, but this field is not commonly used. Access devices, networks, and network routers cannot implement QoS mechanisms because this field does not have the appropriate bits set by the source processor. 5. TCP-Mediated Transmission Rate mechanism [0142]

The IP-centric wireless QoS mechanism can integrate and manage how TCP controls transmission rates. If the TCP mechanism is not managed, the radio QoS mechanism can be overwhelmed and negated by the radio bandwidth factor. Before specifying a specific radio factor that can affect the TCP transmission rate, it is necessary to consider the TCP transmission rate mechanism. [0143]

When packet loss occurs, TCP can "detect" the packet loss and control the transmission rate. Since TCP / IP is primarily created for wired environments with very low unique BERs, such as fiber optic lines, TCP considers packet loss to be due to network congestion rather than due to bit errors. Therefore, TCP reacts to lower the transmit rate, assuming that the transmit rate exceeds the capacity of the network. However, in the case of wireless link segments, packet loss is caused not by congestion but by the primary high inherent BER. This difference cannot be ignored. [0144]

At the start of packet flow, TCP first increases the transmission rate. This is called "slow start mode". The transmission rate can continue to rise until a packet loss or timeout response message is received. TCP then "backs off", that is, reduces the size of the transmit window, significantly reduces the transmit rate, and resends the lost packets in the proper order. TCP can then slowly and linearly increase the transmit rate. This is called congestion prevention mode. [0145]

When a plurality of users share the wireless radio link of the present invention, the medium-specific BER is high, so packet loss may occur frequently and useless TCP retransmission may be performed in the congestion prevention mode. Since wireless bandwidth can be a valuable commodity, it is preferable that the IP-centric wireless QoS mechanism perform packet retransmissions without invoking TCP retransmissions and further "truncating" the transmission rate unnecessarily. This, along with some other factors, is preferably created on the IP-centric wireless medium access control (MAC) layer. One of the functions of the IP-centric MAC layer can be a function to retransmit lost packets locally without sending a signal to TCP and changing the TCP transmission speed unnecessarily. The primary role of the IP-centric wireless MAC layer is to provide shared access to wireless media in an orderly and effective manner. PRIMMA (Proactive Reservation-based Intelligent Multimedia-aware Media) provided by Malibu Networks, Inc., Calabasas, California, MAC layer according to the present invention. The Access) layer can also schedule all packet transmissions over the radio medium based on IP flow type, service level agreement (SLA), and QoS considerations. 6. Avoid TCP congestion in IP-centric wireless systems Network Congestion Collapse, Global Synchronization, and IP-centric wireless TCP congestion avoidance [0146]

Since wireless transmission has a high inherent BER (Bit Error Rate), a problem called congestion discard or global synchronization discard is more likely to occur than in a wired environment. If multiple TCP senders detect congestion due to packet loss at the same time, all TCP senders can enter TCP slow start mode, reduce their transmission window size, and temporarily enter hibernation. After that, multiple senders can try to resend the lost packet at the same time. Since all senders can resume transmission with a rough synchronization, the probability of congestion is high and the same problem may be repeated again. [0147]

In a wireless environment, burst noise can cause packet loss from many IP streams. The TCP transmit rate mechanism of TCP senders can attribute packet loss to congestion and synchronously reduce the transmit rates of senders. When resuming, TCP senders can resume with a rough sync, which can cause real congestion in the wireless link segment. If this behavior is repeated for a period of time, the system performance may become unpredictable. One of the causes is the packet to be dropped and the flooding of system queues that causes useless retransmissions. This gets worse and becomes "competitive", and it will take a few minutes for the system to regain stability. It is clear that this has a detrimental effect on QoS. [0148]

For wired, RED (random early detection) can be used to avoid global synchronization. Global synchronization can be avoided by randomly selecting packets from a randomly selected packet flow before congestion is dropped. It monitors the queue, and when the queue depth exceeds a preset limit, RED fires and activates the TCP sender's transmit rate controller asynchronously. This avoids initial congestion that causes congestion abandonment and global synchronization. [0149]

Instead of simply dropping packets randomly, the priority or type of packets can be taken into account when deciding which packets to drop. Randomly, there is also the ability to determine the probability of discarding a flow by packet priority or type. For wireless systems, select UDP packets for real-time IP flows such as H.323 flows, audio streams with more urgent packet time to live (TTL) parameters, without resending or resetting the TCP rate. WRED (weighted random early detection) can be performed. These IP flows are sensitive to latency and jitter, but not so sensitive to packet loss. [0150]

In a well-designed MAC layer in a wireless environment, packet loss due to BER, which can cause congestion abandonment and global synchronization, is due to RED, packet retransmissions by unwanted TCP senders, and TCP transmission rate resetting. It is best to manage by retransmission of local loss packets according to the present invention, rather than dealing with it. This IP-centric wireless system TCP transmission window manager can detect the status of all packets transmitted via the wireless medium by communicating with the MAC layer. b. Effect of fractal self-similar network traffic characteristics vs. Poisson distribution on network congestion [0151]

Traditionally, it has been thought that network traffic can be represented by a Poisson distribution. When the Poisson distribution is used to simulate the system, the sum of thousands of traffic flows becomes an overall uniform network traffic distribution due to the Poisson distribution. In other words, individual bursts of traffic flow can be "averaged" across the network. We have used the network congestion, burst, and dynamic traffic characteristics of this model to create traditional congestion avoidance strategies, design queue buffer sizes for network devices, and predict traffic and capacity limits. [0152]

Recent studies have illustrated that TCP / IP-based traffic behaves fractally or self-similarly to the network. According to this model, when the bursts of individual traffic flows are summed across the network, the entire network is in a burst state. The burst nature of network traffic flows can be seen across the network, both on a time scale and on a flow scale. This gives great hints for both the design of the IP-centric wireless broadband system according to the present invention and the design of the congestion avoidance strategy for the entire network. This new perspective on network operation has revealed that network router equipment, switch equipment, and transmission equipment are often "technically flawed." Technical deficiencies further exacerbate network congestion behavior. [0153]

The IP-centric wireless system architecture and design tips mentioned above range from queue buffer capacity to local congestion avoidance strategies. Due to the disadvantage of high inherent BERs in wireless systems, the impact of network-wide congestion behavior on local (asexual media channel) congestion avoidance strategies must be properly evaluated and offset. For this reason, congestion avoidance algorithms for IP-centric wireless systems optimize traffic flow by adding new mathematical and engineering considerations that were not known or used by system designers until recently. It is desirable to create it skillfully. [0154]

Taking these considerations into account, the design of an IP-centric wireless system sets it apart from traditional wired system design approaches so as not to severely degrade system performance characteristics. In the traditional design approach of line-centric wireless systems, bandwidth utilization, real-time multimedia quality, and overall system QoS dynamically reduce end-user convenience. 7. Flow control for each application [0155]

In order to achieve the high QoS allowed by end users over multiple data flow ranges, each with different bandwidth, latency, and jitter requirements, IP-centric wireless systems have a wide range of real-time QoS mechanisms. It is desirable that the parameters can be managed. The QoS mechanism needs to be so changeable that the appropriate end user can turn one or more data flows for a particular application on or off in a transparent way. .. This approach differs from other QoS mechanisms that seek to achieve high QoS by establishing a line-centric connection from one end user to another without considering the actual basic application QoS requirements. .. The present invention, which provides a QoS mechanism per application rather than per line, dynamically allocates where the QoS mechanism associated with an individual application type requires, while saving the lack of radio bandwidth. It is possible. B. QoS and IP-centric wireless medium access control 1.PRMMA (Proactive Reservation-based Intelligent Multimedia-aware Media Access) MAC layer [0156]

The PRMMA (Proactive Reservation-based Intelligent Multimedia-aware Media Access) MAC (Media Access Control) layer of the present invention provides an application switching function of an IP-centric wireless QoS mechanism. When other departments of the system determine the nature and QoS requirements for each IP stream, this information is communicated to the PRIMMA MAC layer, which can switch the IP flow of each application to a given destination according to the appropriate priority. 2. PRIMMA IP protocol stack virtual signal [0157]

For IP streams originating from the local user's CPE, the system assigns appropriate QoS mechanism parameters to the IP stream based on application-level information about the nature of the application. For IP streams originating from non-localhost, information about the IP stream is extracted from the packet header and appropriate QoS mechanism parameters are set. Information about the IP stream is communicated "virtually" in the protocol stack model from the application layer (OSI level 7) to the PRIMMA MAC layer (OSI level 2) for bandwidth reservation and application switching. Contrary to the convention that each layer of the protocol stack is independent, this slightly limits the degree of compatibility between the individual layers of the stack, but the advantages are far more disadvantageous in IP-centric broadband access systems. I'm winning. 3. PRIMMA IP flow control and application switching [0158]

Based on each QoS requirement group for each IP application flow of the IP-centric wireless system, the application switches to "proactive" by reserving the appropriate bandwidth via the wireless medium. Radio transmission frames in each direction are constructed in the manner required by the individual QoS requirements of each IP flow. By constructing a wireless transmission frame using QoS requirements, it is possible to obtain optimum QoS performance within the entire range of applications processed by the system. For example, latency and jitter sensitive IP phones, other H.323 compliant IP streams, real-time audio and video streams can be given high priority for optimal placement in wireless transmission frames. .. On the other hand, HTTP (hypertext transport protocol) traffic, such as sending the first web page, can be given a high bandwidth reservation priority for a particular application. Others, such as FTP (file transfer) protocol) Traffic with no latency, jitter, or bandwidth requirements, such as file downloads or email transmissions, can be given low priority for system resources and radio transmission frame placement. 4. PRIMMA TCP transmission rate agent [0159]

The final wireless user is separated from the fast, low BER wired backbone by the slow, high BER radio segment, which is prone to burst errors. TCP / IP traffic across a wireless segment is often subject to packet loss without arbitration, resulting in the aforementioned congestion abandonment and global synchronization. Therefore, the IP-centric wireless system of the present invention uses a TCP transmit rate agent capable of monitoring packet loss in a wireless segment to recreate and transmit an acknowledgment of the lost packet, thereby remote. Can manage the TCP transmission rate function. The PRIMMA MAC layer itself can retransmit packets that have lost radio media. [0160]

The IP-centric wireless TCP transmit rate agent or "auxiliary" can also control the flow of the IP stream, depending on the QoS requirements of the IP flow, if desired. All features of the IP-centric wireless TCP transmit rate agent can be transparent to local and remote hosts and applications. F. Telecommunications network 1. Voice network a. Simple voice network [0161]

FIG. 1A is a block diagram of a standard telecommunications network 100 that provides local telecommunications service providers (LEC) services within one or more local access and transmission areas (LATA). In the telecommunications network 100, the calling party 102 can provide an exchange voice connection to the calling party 110. Figure 1A also shows a private branch exchange (PBX) 112 that allows multiple users to access the LEC service, for example via a leased line. The calling party 102 and the calling party 110 can use a normal telephone device, a key telephone system, a PBX112, or an application running on a host computer. Network 100 can be used as a data connection from the calling party 102, for example, to an ISP (not shown) for modem access. Further, the network 100 is, for example, a private data network. It can also be used to access network) and so on. For example, the calling party 102 can be an employee, for example, operating a laptop computer from a remote location and accessing the employer's dedicated data network via a dial-up modem connection. [0162]

FIG. 1A includes EOs 104 and 108. The EO104 is called an inbound EO because it connects the calling side 102 to the PSTN (Public Switched Telephone Network) equipment. The EO108 is called the outgoing EO because it connects the PSTN equipment to the calling side 110. In addition to inbound EO104 and outbound EO108, PSTN equipment related to telecommunications network 100 has POP (point of presence) 132, 134 locations, for example, one or more IXCs (in the United States) for long-distance traffic. An AT (access tandem) (not shown) can be provided to allow access to a long-distance telecommunications service provider (see Figure 2A). Alternatively, it is self-evident to those skilled in the art that the IXC may also be, for example, CLEC, another ESP (Extended Service Provider), GPOP (International Gateway or International POP), or IP (intelligent peripheral). Is. [0163]

Figure 1A also provides a PBX (private branch exchange) 112 connected to the EO104. The PBX 112 provides a LAN (local area network) 128 consisting of callers 124 and 126, fax 116, client computers 118, and client computers 120 and server computers 122 connected via related modems 130 and related modems 130. You are connected. The PBX112 may be a general class telecommunications device installed at a subscriber site, commonly referred to as a CPE (customer premises equipment). [0164]

Network 100 also includes a CCIS (common channel interoffice signaling) network for setting and tearing down calls. FIG. 1 specifically includes the SS7 (Common Signaling System 7) signal network 114. The signal network 114 will be described later with reference to FIG. 2B. b. Voice network details [0165]

FIG. 2A is a block diagram showing a network 200 that provides LEC and IXC services between subscribers located in different LATAs. Telecommunications network 200 is a more detailed version of telecommunications network 100. The calling parties 102a and 108a are connected to the EO exchanges 104a and 108a, respectively. In other words, the calling side 102a is homed at the inbound direction EO104a of the first LALA, and the calling side 108a is homed at the incoming direction Eo108a of the second LATA. Calls between subscribers between different LATAs are usually long-distance calls transferred by the IXC. Simple US IXCs include AT &amp; T, MCI, and Sprint. [0166]

The telecommunications network 200 includes access tandems (AT) 206 and 208. The AT206 provides connections to POP (point of presence) 132a, 132b, 132c, and 132d. The IXC106a, 106b, and 106c provide connections between POP132a, 132b, and 132c (first LATA) and POP134a, 134b, and 134c (second LATA). CLEC (competitive local exchange carrier) 214 provides an alternative connection between POP132d and POP134d. POP134a, 134b, 134c, and 134d are connected to AT208 in order and to EO108a in the inward direction. The called party 110a receives a call from EO108a, which is the home EO. [0167]

Alternatively, it is self-evident to those skilled in the art that the AT206 may be, for example, CLEC, other ESP (Extended Service Provider), GPOP (International Gateway or International POP), or IP (intelligent peripheral). .. [0168]

The network 200 also has a calling side 102c homed from the CLEC switch 104c. In accordance with the US Telecommunications Act of 1996, CLEC is granted full access within the local RBOC (Bell Regional Bell Operating Company) area of responsibility. i. Fixed wireless CLEC [0169]

The network 200 is also equipped with a fixed radio CLEC209. Fixed wireless CLEC includes, for example, Teligent Inc. (Vienna, Virginia), WinStar Communications Inc., Advanced Radio Telecom Corp., and the BizTel section of Teleport Communications Group Inc. The fixed radio CLEC209 includes a radio transceiver / receiver radio frequency (RF) tower 210 that communicates with the subscriber transceiver RF tower 212 over an RF link. The subscriber RF transceiver tower 212 is connected to the CPE box, PBX112b. The PBX112b connects to a local area network 128b with client computer 120b and server computer 122b, which is connected via caller 124b, 126b, fax 116b, client computer 118b, and associated modem 130b, as well as associated modem 130b. ing. [0170]

Network 200 also comprises a calling party 110a, fax 116a, client computer 118a, and associated modem 130a, cellular communication RF tower 202, and associated cellular subscriber calling party 204 connected to the EO108a, as illustrated further. There is. [0171]

The EO104a, 108a, and AT206,208 form part of the switching hierarchy. The EO104a is also called a Class 5 exchange and the AT208 is also called a Class 3/4 exchange. Prior to the separation of AT &amp; T into RBOC (Bell Regional Bell Operating Company) in accordance with the MFJ (Amendment Consent Decision), the numbers assigned to stations according to the class function of the US PSTN (Public Switched Telephone Network) were in the station class. there were. Station class refers to the functional grade of a telephone office center office switchboard ranked according to transmission requirements and hierarchical relationships with other exchange centers. Class 1 stations are called RCs (Regional Centers) and complete their calls at the highest level stations, the "last base" stations. Class 2 stations are called SCs (Sectional Centers). Class 3 stations are called PCs (Primary Centers). Class 4 stations are called TC (Toll Center) when there is an operator, and TP (Toll Point) when there is no operator. Class 5 stations are EO (end stations), that is, local central office Called (local central station), it is located at the lowest level in local and long range switching and is closest to the last subscriber. Each center handles traffic from one or more lower rank centers. With AT &amp; T's separatist and exchange intelligent software becoming more sophisticated, these classifications have become less robust. As technology provides features closer to the end user, the definition of traditional network hierarchies and switch classes has become widespread. ii. Connectivity with Internet Service Providers (ISPs) [0172]

In addition to providing a voice connection from the calling party 102a to the calling party 110a, the PSTN can provide the calling party 102 with a data connection with an ISP (such as client 118b). [0173]

The network 200 can also include an Internet Service Provider (ISP) (not shown) with a server computer 122 connected to a data network 142 (discussed below with reference to FIG. 1B). The Internet, as is already known, is a global network consisting of multiple large networks connected by data links. These links include, for example, ISDN (Integrated Services Digital Network), T1, T3, FDDI, and SONET links. Alternatively, a dedicated network in which multiple LANs and / or WANs are interconnected, such as an intranet, may be referred to as the Internet. The ISP can provide Internet access services for subscribers such as Client 118b. [0174]

When establishing a connection with an ISP, the client 118b can use the host computer connected to the modem (modulation demodulator) 130b. The modem modulates the data from the host computer into a format that sends it to the LEC equipment (conventional analog format). LEC equipment typically converts the input analog signal to digital format. In one embodiment, the data is converted to point-to-point (PPP) format. (PPP is a well-known protocol that allows a computer to establish a connection to the Internet using a standard modem.) For those skilled in the art, transmission control programs, the Internet Protocol (TCP /) are other formats. IP) Packet Format, User Datagram Protocol, Internet Protocol (UDP / IP) Packet Format, Asynchronous Transmission Mode (ATM) Cell Packet Format, Serial Line Interface Protocol (SLIP) Protocol Format, Point-to-Point (PPP) Protocol Format, Point Two-point tunneling protocol (PPTP) format, NETBIOS extended user interface (NETBEUI) protocol format, Appletalk protocol format, DECnet, BANYAN / VINES, IPX (internet packet exchange) protocol format, Internet Control Message Protocol (ICMP) protocol format, etc. It is self-evident that it is available. iii. Communication link [0175]

Figures 1A, 2A and other drawings in this document describe network nodes, or lines that represent communication lines or logical connections between systems, physically performed by a telecommunications carrier's equipment. Telecommunications carrier equipment includes lines and network nodes between the lines, which include, for example, digital access cross-connect systems (DACS), repeaters (regenerators), tandems, copper wires, and fiber optic cables. Lines such as. It is self-evident to those skilled in the art that one or more telecommunications equipment will be connected and used as an alternative communication line. Further, the telecommunications carrier defined here includes, for example, LEC, CLEC, IXC, ESP (extended service provider), international service provider such as GPOP (global point of presence), IP (intelligent peripheral) and the like. [0176]

The EO104a and AT206 are connected by a trunk. The trunk connects the AT to the EO. The trunk can also be called an IMT (inter machine trunk). The trunk connecting AT208 and EO108a may also be IMT. [0177]

With reference to Figure 1A, the EO104 and PBX112 can be connected via a dedicated dial tone line. The leased line can also connect an ISP (not shown) to, for example, the EO104. The dial tone leased line can be connected to the modem bay or access converter device at the ISP. Examples of leased lines include channelized T1 (channelized T1) or ISDN (Integrated Services Digital Network) PRI (Primary Group Interface). ISPs can also connect to the Internet via pipes or dedicated communication equipment. The pipe may be a dedicated communication facility. The leased line can handle sending and receiving data modem traffic to and from your ISP. [0178]

The trunk can handle switched voice traffic and data traffic. For example, the trunk can contain DS1 to DS4 digital signals transmitted over the T1 to T4 carriers. Table 2 lists common carriers and their digital signals, number of channels, and bandwidth capacity.

[Table 2]

<img file="JP2003521138A_D0007.tif" /> 【0179】

Alternatively, the trunk can include optical carriers such as OC-1, OC-3, and the like. Table 3 lists common optical carriers and their respective synchronous transmission signals (STS), ITU (International Telecommunication Union) names, and bandwidth capacities.

[Table 3]

<img file="JP2003521138A_D0008.tif" /> 【0180】

As already mentioned, a leased line is a connection that can carry data modem traffic. The leased line can be a direct channel between two designated points that the customer uses exclusively. Leased lines are also sometimes referred to as rental lines. In one embodiment, an ISDN / primary group interface (PRI) connection is used for the leased line. An ISDN PRI connection can include a T1 signal channel (called a data channel or D channel) and 23 channels used as bearer channels or B channels. (A bearer channel is a digital channel that transmits voice and data information.) Even if multiple ISDN PRI lines are used, the signals of all lines can be transmitted via one D channel, so the remaining lines. Can be used as a bearer channel. iv. Telecommunications traffic [0181]

Telecommunications traffic can be sent and received from the network nodes of telecommunications carriers. Telecommunications carriers include, for example, LEC, CLEC, IXC, ESP (Extended Service Provider) and the like. In one embodiment, this traffic can be received from a network node that is, for example, a Class 5 exchange such as EO104a or a Class 3/4 exchange such as AT206. Alternatively, the network system can also include, for example, CLEC, or other ESP (Extended Service Provider), International Gateway, POP (point-of-presence), GPOP, IP (intelligent peripheral). [0182]

The voice traffic refers, for example, an exchange voice connection between the calling side 102a and the called side 110a. It should be noted that this is a point-to-point dedicated path, that is, bandwidth is allocated regardless of whether it is in use or not. When an exchange voice connection is established between the calling side 102a and the called side 110a, it connects from IXC106a to AT206 via the IXC network such as AT208, then to EO108a, and to the called side 110a via the trunk. Will be done. In other embodiments, the AT206 or IXC106a may also be, for example, CLEC, another ESP (Extended Service Provider), an international gateway, a GPOP (global point-of-presence) or an IP (intelligent peripheral). [0183]

The calling side 102a may be a computer that has data connected to the server via a voice network. Data traffic refers, for example, to the data connection between the calling party 102a (using a modem) and the server 122b (which may be part of an ISP). A data connection is established, for example, between the calling party 102a and EO104a, then to the PBX112b via the AT206, CLEC214, and then the fixed wireless CLEC209 link, and to the modem 130b associated with the server 112b. I can go. c. Signal network [0184]

FIG. 2B is a diagram illustrating the signal network 114 in more detail. The signal network 114 is an independent network used to set up, discard, and manage calls between the calling party 102 and the calling party 110. The signal network 114 in this example is a common line signal system 7 (SS7) network. The common line signal network 7 includes SSP (service switching points) 236,238, 240, 242, STP (signal transfer points) 222, 224, 226, 228, 230, 232, and SCP (service control point) 234. [0185]

In the SS7 network, SSP is part of the backbone switch that provides SS7 functionality. The SSP may be, for example, a combination of a voice switch and an SS7 switch, or a computer connected to the voice switch. The SSP uses primitives to communicate with the switch and generate packets to be sent over the SS7 network. [0186]

In SS7 signal network 114, EO104a, 108a, and AT206,208 can be represented by SSP236, 238, 240 and 242, respectively. Therefore, the connections between EO104a, 108a and AT206, 208 (indicated by the dotted lines) can be represented by connections 254, 256, 258, 269, respectively. These types of links are described below. [0187]

The STP acts as a router for the SS7 network and is usually provided as an auxiliary device in place of the switchboard. The STP forwards messages from the outgoing SSP to the source SSP. From an architectural point of view, STPs can usually be "matted pairs" to share resources with redundancy in the event of congestion or failure (automatic load sharing). To do). As shown in Figure 2B, STPs can be arranged according to hierarchical levels to forward signal messages hierarchically. For example, the paired STP222 and 224 and the paired STP226 and 228 are at the first level, and the paired STP230 and 232 are at the second level. [0188]

SCP provides database functionality. Use SCP to provide extensions such as forwarding special service numbers (such as 800 and 900), storing information about subscriber services, checking calling cards and fraud prevention, and providing Enhanced Intelligent Network (AIN) services. can do. SCP234 is connected to a pair of STP230,232. [0189]

In the SS7 network, there are unique links between different network elements. Table 4 describes the definitions of common SS7 links. [0190]

As shown in Fig. 2B, the paired STPs are connected by a C link. For example, the STP222 and 224 pairs, the STP226 and 228 pairs, and the STP230 and 232 pairs are connected by C-links (not marked). The pairs of SSP236 and 238 and the pairs of SSP240 and 242 are connected by F-links 262 and 264, respectively. [0191]

The pairs of STP222, 224 and the pairs of STP226, 228 at the same hierarchy level are connected by B-links 270, 272, 244 and 282. Pairs of STP222, 224 and pairs of STP 230, 232 at different hierarchy levels are connected by D-links 266, 268, 272 and 274. Similarly, pairs of STP226, 228 and pairs of STP230,232 at different hierarchy levels are connected by D-links 278, 280, 246 and 248. [0192]

The pairs of SSP236 and 238 and the pairs of STP222 and 224 are connected by A links 254 and 256. The pairs of SSP240 and 242 and the pairs of STP226 and 228 are connected by A-links 258 and 260. [0193]

The SSP236, 238 pair is also connected to the STP230, 232 pair by an E-link (not shown). Finally, the STP230,232 pair is connected to SCP234 by A-link 250,252. [0194]

For a more detailed explanation of the SS7 network topology, refer to Russel, Travis, Signaling System # 7, McGraw-Hill, New York 10020, ISBN 0-07-054991-5, which is referenced in this document.

[Table 4]

<img file="JP2003521138A_D0009.tif" /> d. Call flow with SS7 signal [0195]

When starting a call on the SS7 telecommunications network, the calling party first dials the calling party's phone number using the phone connected to the incoming EO switch. The dialed phone number is sent from the phone to the SSP of the inbound EO of the LEC (local telecommunications service provider) in the calling area. The SSP first begins processing internal routing conventions based on whether it meets certain criteria. The SSP then initiates a signaling message to another EO or Access Tandem (AT) as needed. The signal information is sent from the SSP to the STP, and the signal is transferred between the inbound EO and the called party EO or the outgoing EO. The outgoing EO has a port specified by the calling party's telephone number. The call is set up as a direct connection between EOs via a tandem switch if there is no direct trunking or if the direct trunking is full. If the call is a long-distance call, that is, if the calling and calling parties are in different LATAs (local access transmission areas), the calls are connected via IXC (Long Distance Telecommunications Service Provider). Such long-distance calls are commonly referred to as inter-LATA calls. LEC and IXC are collectively called PSTN (Public Switched Telephone Network). [0196]

The passage of the Telecommunications Act in 1996, which allows competition in the regional telephone services market, allowed CLEC to compete with ILEC to provide regional telecommunications services. However, this competition has not provided the bandwidth needed to handle large amounts of voice and data communications. This is due to the limitation of circuit switching technology, which limits the bandwidth of the equipment used by LEC, and the huge capital investment required. e. Circuit switching [0197]

Circuit switching dedicates a channel to a call for the duration of that call. For this reason, circuit switching requires a large amount of exchange bandwidth in order to process a large amount of voice calls. This problem is complicated by voice lines that carry out data communication via devices designed to handle voice communication. i. Time Division Multiplexing (TDM) Circuit Switching [0198]

TDM circuit switching creates a full-time connection or a leased line for the duration of the connection between two connected devices. TDM divides the bandwidth into fixed time slots. In this way, it is possible to generate a plurality of time slots, each of which can use its own defined capacity. Each device connected by the TDM network is assigned a fixed portion of the bandwidth using one or more time slots according to the required speed. When the device is in transmit mode, the data is simply placed in the time slot, with no overhead such as processing or conversion. Therefore, TDM is a transparent protocol for transmitted traffic. Ultimately, however, when the device is not transmitting data, the time slots are free and bandwidth is wasted. High-speed devices on the network may slow down while waiting for data transmission, but the idle and standby capacity cannot be allocated to this high-priority device during transmission. Data bursts are becoming quite commonplace in today's society due to demand, but TDM does not adequately support these data bursts. 2. Data network [0199]

As an example, FIG. 1B shows a network 148 including workstations 144 and 146 connected to the data network 142. The data network 142 operates as a wide area network (WAN) that connects a plurality of local area networks (LANs). As an example, network 148 includes a plurality of host computers connected by a hub such as a NIC (Network Interface Card) and an Ethernet® hub, such as client workstation 137 and server 136. The LAN is connected to the data network 142 by the network router 140, which allows the date traffic to be forwarded from the client 138 and the server 136 to the workstations 144 and 146. a. Packet switching [0200]

Unlike the voice networks 100 and 200, which carry traffic over circuit-switched connections, as shown in FIGS. 1A and 2A, data network 148 carries traffic by packet switching. [0201]

Currently, the Internet, intranets, and similar public and dedicated data networks that interconnect computers generally use packet-switched technology. Packet switching is a more efficient telecommunications channel than circuit switching. Packet-switched networks carry packets of information that can contain various types of data, such as digitized voice, data, and video. Packet switching allows a large number of different calls to share a communication channel, rather than dedicating a channel to a single call. In voice calls, for example, digital voice information is transmitted between callers for 60% of the time and may be silent for the remaining 40% of the time. For circuit-switched connections, voice calls are tied to communication channels where 50% of the bandwidth is not used for silence. For data calls, information may be transmitted between two computers using 10% of the time. For data calls, 90% of the bandwidth on the channel is unused. In contrast, in packet-switched connections, voice calls, data calls, and other possible call information are all transmitted over the same channel. [0202]

Packet switching divides a medium stream into smaller pieces, for example called packets, cells, or frames. Each packet is then encoded with the address information of the destination to be transmitted and can be transmitted over the network. The packet is received at the destination, assembled in its original format, and transmitted to the recipient. This process can be performed according to one of the important communication protocols commonly called IP (Internet Protocol). [0203]

In packet-switched networks, there is no fixed continuous physical connection between the sender and the receiver. Packets transmitted from many different calls share network bandwidth with other transmissions. The packet is then simultaneously sent to the destination via many different routes. Packets arriving at the receiving end can be reassembled. This allows telecommunications bandwidth to be used much more efficiently than circuit switching. b. Router [0204]

The data network 142 includes a plurality of network routers 40. Network routers can be used to transfer information between multiple networks. A router acts as an interface between two or more networks. A router can find the best path between two networks, even if there are several different networks between the two networks. [0205]

A network router can have a table listing various network domains. The domain is a local area network (LAN) can be considered as a wide area network (WAN). Information can be transferred between multiple LANs and / or WANs via network routers. When the router receives the packet, it determines the destination domain of the packet from the destination address in the header of the packet. If the router is not directly connected to the destination domain, the router can forward the packet to the router's default router, that is, a router higher in the router hierarchy. Because each router has a connected default router, packets can be sent to the destination domain through multiple consecutive routers and to the destination host with the final destination address of the packet. it can. c. Local Area Network (LAN) and Wide Area Network (WAN) [0206]

A local area network (LAN) can be thought of as multiple host computers that are interconnected via a network interface card (NIC) on the host computer. The NICs are connected, for example, via copper wires so that the host computers can communicate with each other. Examples of LANs include Ethernet bus networks, Ethernet switch networks, Token Ring networks, FDDI (fiber digital data interconnect) networks, and ATM networks. [0207]

A wide area network (WAN) is a network that connects host computers over a wide area. A network interface that interconnects the LAN and WAN is required so that the host computer on a particular LAN can communicate with the host computer on another LAN or WAN. The router described above is an example of a network interface. [0208]

A network designed to interconnect multiple LANs and / or WANs is called the Internet (internet: "i" is lowercase). The Internet (where "i" is lowercase) can transfer data between multiple networks, including LANs and WANs. Communication between a host computer on one LAN and a host computer on another WAN can be performed, for example, by using the Internet Protocol (IP) protocol. When you use the IP protocol to assign a single IP address to each host computer on each network, packets are transmitted over the Internet to other LANs and / or WANs connected to the Internet. Will be possible. The Internet (i is lowercase) has routers interconnected to two or more networks. [0209]

The "Internet" ("I" is capitalized) is the global Internet interconnected with all networks around the world. The Internet (I is uppercase) consists of a global network of computers interconnected via the Internet Protocol (IP) protocol. [0210]

An "intranet" is the Internet (where "i" is lowercase), a dedicated network that uses Internet software and Internet standards. An intranet can be reserved for use by a person authorized to use the network. d. Switching and routing [0211]

Routing is performed on medium-level network architectures with protocols such as IPX or TCP / IP. Switching is performed at a lower level, Layer 2 of the OSI model, the Medium Access Control (MAC) layer. e. TCP / IP packet-centric data network vs. ATM line-centric data network [0212]

Asynchronous transmission mode (ATM) is a line-centric data network that exchanges fixed-sized cells. ATMs execute virtual circuits (VCs), virtual paths (VPs), and transmit paths (TPs). Line-centric networks, such as ATMs, set up virtual circuits between source and destination nodes that provide QoS by dedicating virtual circuits to individual traffic types. [0213]

Some networks are packet-centric networks. Unlike line-centric networks, packet-centric networks do not use leased lines to transfer packets. TCP / IP packets user data and sends it between various systems on an IP network. When a large file is sent to the protocol stack, the IP function divides the data into segments and then into packets. The packet is attached with a header and transmitted to the data link. This data is routed and switched at the IP (ie network) layer. IP is, in a sense, a dump protocol. When the packet is ready to be sent over the medium, IP does not explicitly route the call over a particular channel. Instead, add a header to the packet and let the network process it. Therefore, the outward packet is transmitted from the source to the destination via various routes. This means that packets are in the form of datagrams, rather than being numbered in sequence like other protocols. IP does its best to send packets to the destination network interface, but whether the data arrives, whether the data is sent error-free, and which nodes the data passes through are accurate in their order and order. There is no guarantee that they will be involved or that they will warn the sender if something goes wrong during the transmission. IP has a header information feature that allows a certain number of "hops", or so-called "expirations," on the network, because when packet is IP routed, the transmitted packet may loop over the network. .. IP has a counter function that decrements each time a packet passes through a network node so that undelivered "packages" do not continue to circulate in the network. The node discards the packet whose counter has expired. Along with IP, TCP controls to ensure reliable transmission and transmission of data streams. TCP at the transmitting end Attaches a byte count header to the information sent to the IP protocol layer and encapsulates it as part of the packet. At the receiving end, when a packet is received, it sorts the packet to ensure accuracy. If all IP flows are not received correctly, a byte count acknowledgment or negative response is sent back to the sender, prompting the sender to resend the bytes needed to fill the rest of the packet flow. TCP also buffers additional packets until the negatively answered packet is retransmitted. 3. Video network [0214]

Figure 1C shows a traditional video network 150, such as a cable television (CATV) network. The video network 150 can include various video captures, distribution links, and a video network 160 connected to a video output monitor. Video input devices can include, for example, conference cameras, 154, 159. The video output device can include, for example, televisions 152, 156. The video network 160 includes a head end (serving end of the cable), a coaxial cable television (CATV), and an NTSC (national television standard code) tuner device that multiplexes various video signals. be able to. Standard cable systems have a huge amount of available bandwidth. [0215]

It should be noted that CATV is a wireless communication method. The frequencies of many video signals are distributed simultaneously along the cable. A TV tuner selects a particular channel by tuning to a particular frequency or "frequency band". [0216]

Cable TV CATV video networks often have only one physical cable, but cables can have multiple channels at the same time. This is possible by sharing the frequency spectrum of the cable and using frequency division multiplexing (FDM) to allocate multiple frequency ranges to multiple channels. Broadband cable communication systems behave exactly like CATV systems. This FDM technology counter does not divide the cable into frequency bands, but divides it into time slots by time division multiplexing (TDM). TDM allows each video transmitter to acquire the entire frequency band of the cable, but only for a very short period of time. The cable is now capable of transmitting up to 750MHz. FDM technology can be used to split a channel into multiple dedicated logical channels. Technological innovation has made it possible to perform time division multiple access (TDMA) within the FDM channel. [0217]

A cable system can multiplex two independent dimensions to create multiple data channels on a single cable. Channels are separated by FDM, and one frequency band can be shared by multiple users by TDMA. The most common method of applying the TDMA access method over a wide band is CSMA / CD developed by XEROX for Ethernet. [0218]

A single cable is used to enable simultaneous bidirectional transmission with a midsplit. Another method is to use a dual cable system for simultaneous bidirectional transmission. [0219]

Broadband is basically an analog signal method. For example, since a camcorder is also an analog device, signals from the camcorder (or recorder) can be transmitted directly over a wideband cable channel in the red / green / blue (RGB) format. G. Voice / Data / Video Network Aggregation [0220]

Recognized for the efficiencies inherent in packet-switched data networks such as the Internet, the main focus has recently been on digitizing voice, data, video and other information for transmission over aggregated packet-switched data networks. There is. To provide high quality of service (QoS) to end users, data networks strive to provide various types of information in a timely manner with appropriate bandwidth and convenience that end users can tolerate. [0221]

As an example, Figure 2C shows network 286, which transmits voice, data, and video traffic over a data network. Network 286 has a calling party 102b homed at EO104b, which is linked to a telephone gateway 288b. Network 286 also has a called party 110c homed at EO108c, which links to the telephone gateway 288c. The EO104b, 108c and the telephone gateways 288b, 288c can be linked to the signal network 114. The telephone gateways 288b and 288c can also be connected to the data network 142 via routers 140b and 140c, respectively. [0222]

As shown in FIG. 1C, the telephone gateways 288b, 288c can be used to packetize voice traffic and signal information into a suitable format and transmit it over the data network 142. Those skilled in the art will appreciate that the telephone gateways 288b, 288c are equipped with a variety of computer devices designed for call control, configuration and discarding. Voice calls transmitted over the data network are, for example, VoP (voice over packet), VoD (voice over data), VoIP (voice over internet protocol), VoATM (voice over asynchronous transfer mode), VoF (voice over frame). And so on. [0223]

Examples of telephone gateways 288b and 288c include MGCP (media gateway protocol) compliant gateways manufactured by vendors such as Luent (Parsippany, New Jersey) and CISCO (Palo Alto, California). It should also be noted that there are other network devices capable of transmitting VoIP, such as softswitches available from SoftSwitch Consertium member companies such as Level 3 Communications of Lousvile (Colorado). [0224]

The network 286 also includes other devices connected to the network 142 as shown. First, the H.323 compliant video conferencing system 289 is equipped with a illustrated camera of 154g, a television of 152g, and a router of 140g. Second, a local area network (LAN) 128a with a client workstation 138a and a server 136a connects to the data network 142 via a network router 140a. Similarly, LAN128f with client workstation 138f and server 136f is connected to data network 142 via network router 140f. [0225]

The data network 142 can transmit a packet of information from a source to a destination connected to the data network 142 via a network routing device. For example, data network 142 can send voice and data traffic from telephone gateway 288b to telephone gateway 299c by transmitting Internet Protocol (IP) packets. Data network 142 represents a packet-centric data network. A well-known data network is the global Internet. Other examples include dedicated intranets, packet-switched networks, Frame Relay networks, and asynchronous transmission mode (ATM) circuit-centric networks. [0226]

As another embodiment, the data network 142 may be an IP packet switching network. For example, a packet-switched network such as an IP network does not require a dedicated line between a calling end and a calling end in the packet-switched network, unlike a circuit-switched network. Instead, packet-switched networks divide messages into small pieces called packets of information. The packet divided in this way is encapsulated with a header that specifies the destination address that is the destination of the packet. When a packet-switched network receives a packet, it then forwards the packet to the destination indicated by the destination address in the packet header. [0227]

Routers 140a, 140b, 140c, 140d, 140e, 140f, 140g can be connected to each other by physical media such as optical fiber link connection and copper wire connection. Routers 140a to 140g transfer information according to a routing protocol and communicate with each other. [0228]

Data network 142 is run by IP networks, ATM virtual circuit-centric networks, Frame Relay networks, X.25 networks, and other types of LANs and WANs. Other networks can interoperate with data networks 142, such as FDDI, Fast Ethernet (Fast Ethernet®), or SMDS packet-switched networks. Frame Relay networks and ATMs are connected, line-centric services. SMDS (switched multi-megabyte data service) is a connected mass packet service that can provide speeds up to 45 Mbps. 1. Data network example Asynchronous Transfer Mode (ATM) [0229]

ATM is a high-bandwidth, low-latency, multi-network technology that uses fixed-size cells. Bandwidth capacity is segmented into 53-byte cells with header and payload fields. ATMs use fixed-length cells because they are easier to exchange than variable-length packets in terms of hardware, and transmission will be faster in some environments. [0230]

The ATM environment sets up virtual circuits in a line-centric way. Therefore, ATM uses SAR (segmentation and resequencing algorithm) to segment variable-length IP packet flows into fixed-size cells. [0231]

An ATM cell can have a 48-byte payload field and a 5-byte header that identifies the cell's so-called "virtual circuit". ATMs are considered suitable for fast combinations of voice, data and video services. Currently, ATM access is possible at speeds of 622 Mbps and above. The maximum speed that ATM can achieve is doubling every year. [0232]

ATM is defined by protocols standardized by the ITU-T (International Telecommunication Union), ANSI (American National Standards Institute), ETSI, and ATM Forums. ATM consists of multiple building blocks such as transmit paths, virtual paths, and virtual channels. Asynchronous Transfer Mode (ATM) is a cell-based exchange multiplexing technology designed as a line-centric transfer mode for a wide range of telecommunications services. ATM can also be applied to LAN and dedicated network technologies, as specified by the ATM Forum. [0233]

ATM handles connected traffic directly or through the adaptation layer, or uses the adaptation layer to handle disconnected traffic. ATM virtual connections can work with either CBR (Constant Bit Rate) or VBR (Variable Bit Rate). ATM cells sent to an ATM network contain a small header containing information that establishes a virtual circuit-centric connection from the caller to the destination. All cells are sequentially transferred over this virtual connection. ATM provides both PVC (fixed connection) and SVC (selective connection). ATM is asynchronous. This is because the transmitted cells do not have to be periodic because the time slot is required to be in synchronous transfer mode (STM). [0234]

ATM uses a method of adding a header field to the end of a fixed-length payload. The AMT header identifies the virtual channel (VC). Therefore, the time slot can be used by a host with data ready for transmission. Empty or idle cells are sent if no host is ready to send. [0235]

ATM has become the standard for network architectures that define multiple exchange methods. SONET (synchronous optical network) provides the physical basis for ultra-fast transmission. ATM can also provide separate virtual circuits for each traffic type, depending on latency and loss performance, and support multiple quality of service (QoS) classes for different application requirements. ATM can also support LAN-like access to the available bandwidth. [0236]

Cells are mapped to physical transmission paths such as North American DS1, DS3, SONET; European E1, E3, E4; ITU-T STM standards; and local fiber and electrical transmission payloads. All information is multiplexed and exchanged by these fixed-length cells in the ATM network. [0237]

The ATM cell header field consists of six parts that identify the cell type and priority. ATM cell headers include GFC (general flow control), VPI (virtual path identifier), VCI (virtual channel identifier), PT (payload type) CLP (call loss priority) and HEC (header error checking). VPIs and VCIs identify destinations and are only important locally. GFC lets the multiplexing device control the rate of the ATM terminal. The PT indicates whether the cell has user data, signal data, or maintenance information, and the CLP indicates the relative importance of the cell. During periods of congestion, lower priority cells are discarded before higher priority cells. HEC detects and corrects header errors. [0238]

The ATM cell payload field passes between networks as is, without checking or correcting errors. ATM relies on higher layer protocols to perform payload error checking and correction. For example, the error correction function can be performed using Transmission Control Protocol (TCP). Since the cell size is fixed, ATM exchange and multiplexing are simple and can be performed at high speed. [0239]

In the case of ATM, long packets are divided into many fixed-length cells, so unlike other packet-switched networks, long packets do not lag behind short packets. This feature allows ATMs to carry CBR traffic, such as voice and video, in the same network along with VBR data traffic, which can potentially have very long packets. [0240]

ATMs exchange traffic, segment fixed-length cells, multiplex cells into a single bitstream, and send them over a physical medium. Different types of traffic, such as voice, video, and data traffic, can be sent over the ATM network. Video and audio traffic is very time sensitive and delays do not have significant fluctuations. On the other hand, data can be transmitted in either connected or disconnected mode. In either case, the data is not as sensitive to delay as voice and video traffic. Data traffic, such as spreadsheet data, requires transmission accuracy. Therefore, ATMs have traditionally had to distinguish between voice, video and data traffic. Voice and video traffic must be prioritized and transmission delays must always be within a limited range. But data traffic must also be low loss. In an aggregated data network, data traffic can also carry voice traffic, which is time-dependent. In one example of running an ATM, multiple types of traffic can be combined with virtual circuits (VCs) assigned to different data, voice, and video traffic through one ATM virtual path (VP). is there. [0241]

The transmission path can contain one or more VPs. Each VP can contain one or more VCs. Therefore, a plurality of VCs can use one VP as a trunk. The exchange can be performed at the transmission path, VP or VC level. [0242]

The ability of ATMs to switch to the virtual channel level is similar to the behavior of a private branch exchange (PBX) or telephone switchboard. In the case of a PBX switch, each channel in the trunk can be replaced individually. Since it is similar to a telephone exchange, a device that executes VC communication is generally called a VC exchange. A device that performs VP communication is commonly referred to as a VP cross-connect because it resembles a transmit network. These names are meant to be easy to understand and their meanings cannot be interpreted literally. ATM cell exchanges need not be limited to VC-only exchanges and VP-only cross-connects. [0243]

At the ATM layer, the user can choose between a virtual path connection (VPC) or a virtual channel connection (VCC). Virtual path connections (VPCs) are exchanged based solely on the value of the virtual path identifier (VPI). Within the VPI, VCC follows the same route, so VPC users can transparently assign VCC. The virtual channel connection (VCC) is exchanged according to the value of the VPI and the virtual channel identifier (VCI). [0244]

Calls can be transmitted over the network using both VPI and VCI. Note that the VPI and VCI values must be unique for the same outbound path (TP). [0245]

It should be noted here that the network 142 may be an ATM network as well as other data type networks including various packet-switched data type networks. b. Frame Relay [0246]

Alternatively, the data network 142 may be a Frame Relay network. It is self-evident to those skilled in the art that the Frame Relay network can be used as the data network 142. Data, not ATM cells, can be transmitted in frames. [0247]

Frame Relay is a packet-switched protocol that is used on WANs and is becoming more common for remote LAN-to-LAN connections. [0248]

Previously, Frame Relay access speeds were less than about 1.5 Mbps. Today, so-called "fast" Frame Relay at about 45 Mbps is offered. This speed is still slow compared to other technologies such as ATM. [0249]

The Frame Relay service employs a form of packet switching similar to a streamlined version of the X.25 network. Packets are in the form of variable length frames. The advantage of this method is that Frame Relay networks can accommodate data packets of various sizes associated with virtually negative data protocols. Frame Relay networks are relatively protocol independent. In the example of the frame-frame network of the data network 142, the process of the conventional protocol convention is not executed, so that the exchange can be performed faster and cheaper than some alternative networks. Frame Relay is also designed for the reliable lines currently available and is faster than traditional X.25 networks because it performs relatively less stringent error detection. c. Internet Protocol (IP) [0250]

In some embodiments, network 142 may be an Internet Protocol (IP) network over an ATM network. It is self-evident to those skilled in the art that an Internet Protocol (IP) network via various other data link layer networks such as Ethernet can be used as a data network network 142. Data can be transmitted in variable-length IP datagram packet-centric packets segmented by TCP instead of fixed-length ATM line-centric cells. IP data networks can be superior to some physical networks, such as SONET fiber optic networks. 2. Virtual Private Network (VPN) [0251]

A virtual private network (VPN) looks like a leased line when used, but in reality it is a wide area communication network provided by a telecommunications carrier, including trunks shared by all customers on public networks. Is. VPNs can provide services over wired networks, but wireless networks can provide services as well. A VPN can configure a dedicated network within a public network. [0252]

VPNs allow carriers to provide their customers with secure, guaranteed, long-range bandwidth for WANs. These VPNs can generally be used by selecting either Frame Relay or SMDS (switched multi-megabyte data service) protocols. These protocols logically define groups of users on the network, regardless of their physical location. ATMs are the preferred VPN protocol because companies that handle complex applications require high reliability and high bandwidth. VPNs use ATMs to provide such corporate networks with WAN-like virtual security and QoS corporate networks designed with leased lines. [0253]

The Internet has created a much cheaper virtual private Internet instead of a VPN. VPI (Virtual Private Internet) allows businesses to connect to different LANs over the Internet. Users can build a secure shared intranet with VPN-style network permissions and encryption by installing software alone or a combination of hardware and software. VPIs typically use a browser-based administration interface. 3. H.323 video conferencing [0254]

A brief overview of H.323 recommendations for video conferencing. The H.323 standard is the basis for, for example, audio, video, and data communication over IP-based networks, including the Internet. Multimedia products and applications from multimedia vendors that comply with H.323 recommendations allow users to communicate without worrying about compatibility issues. H.323 is expected to be the basis for future LAN-based product multimedia applications. [0255]

H.323 is a recommendation under the International Telecommunication Union (ITU) that discusses standards for multimedia communications over local area networks (LANs) that do not guarantee quality of service (QoS). This network is dominant in today's enterprise desktops, including IPX over Ethernet, packet-switched TCP / IP, high-speed Ethernet, and Token Ring network technology. Therefore, the H.323 standard is a key building block for new, widespread, co-produced, LAN-based applications for multimedia communications. [0256]

The H.323 specification was approved by ITU Research Group 16 in 1996. Version 2 was approved in January 1998. The H.323 standard is extensive, including stand-alone equipment, embedded computer technology, point-to-point conferencing, and point-to-multipoint conferencing. H.323 also covers call control, multimedia management, bandwidth management, and interfaces between LANs and other networks. [0257]

H.323 is part of a set of communication standards that enable video conferencing over a range of networks. The communication standard for this series is called H.32X and includes H.320 and H.324. H.320 and H.324 handle ISDN and PSTN communications, respectively. [0258]

The H.323 architecture defines four key components for network-based communication, including terminals, gateways, gatekeepers, and MCUs (multipoint control units). [0259]

A terminal is a client-side endpoint in a LAN that provides real-time, two-way communication. All terminals support voice communication. Video and data are optional. H.323 defines the modes of operation required for different audio, video, and / or data terminals to work together. H.323 is the standard for next-generation Internet telephones, audio conferencing terminals, and video conferencing technology. [0260]

H.323 terminals also support the H.245, which is used to negotiate channel usage and functionality. Three other components are required. Q.931 for call signals and call settings, which is a component called RAS (Registration / Administration / Status). Used for communication with gatekeepers, it supports RTP / RTCP for ordering audio and video packets. [0261]

Optional components of the H.323 terminal are the video codec, T.120 data conferencing protocol, and MCU functionality. [0262]

The gateway is an optional component of the H.323 conference. H.323 gateways can provide many services, the most common of which is the ability to convert between H.323 conferencing endpoints and other terminal types. This feature includes conversion between transmission formats (H.225 to H.221) and communication procedures (H.245 to H.222). In addition, the gateway also converts between audio and video codecs, setting and erasing calls on both the LAN side and the switched network side. [0263]

In general, the purpose of H.323 gateways is to reflect the characteristics of LAN endpoints on SCN endpoints and the characteristics of SCN endpoints on LAN endpoints. The primary use of gateways is to establish links with analog PSTN terminals, establish links with remote H.320 compliant terminals over ISDN-based switched telephone networks, and remote H.324 compliant over PSTN networks. It is thought that it is in the establishment of a link with the terminal of. [0264]

Since the endpoint can communicate directly with other endpoints on the same LAN, the gateway is not necessary if it does not connect to other networks. The terminal communicates with the gateway using the H.245 and Q.931 protocols. [0265]

By using a suitable transcoder, the H.323 gateway 5806 can support H.310, H.321, H.322, and V.70 compliant terminals. [0266]

Many gateway functions are left to the designer. For example, the actual number of H.323 terminals that can communicate via a gateway is not specified in the standard. Similarly, the number of SCN connections, the number of independent conferences supported at the same time, audio / video / data conversion capabilities, and multiple comprehensive features are also left to the manufacturer. By incorporating H.323 gateway technology into the H.323 specification, the ITU positions H.323 as a means of bringing together standards-based conference endpoints. [0267]

The gatekeeper is the most important component that H.323 can use. It acts as a central point for all calls in the zone and provides call control services to registered endpoints. In many ways, the H.323 gateway acts as a virtual switch. [0268]

The gatekeeper performs two important call control functions. The first call control function is address translation from LAN aliases for terminals and gateways to IP or IPX addresses, as defined in the RAS specification. The second call control function is bandwidth management, which is also specified in the RAS specification. For example, suppose the network manager specifies a threshold for the number of simultaneous meetings on the LAN. Gatekeepers can be confused when establishing further connections once a connection reaches a threshold. This will limit the entire conference bandwidth to a portion of the available bandwidth and dedicate the rest of the functionality to email, file transfer, and other LAN protocols. Terminals, gateways, and multiple control units that can be managed by a single gatekeeper are all collectively called the H.323 zone. [0269]

An optional but valuable feature of the gatekeeper is the H.323 call routing feature. Calls can be controlled more effectively by setting the call through the gatekeeper. Service providers need this feature to charge for calls placed in the network. This service can also be used to reroute a call to another endpoint when the called endpoint (EO) is unavailable. In addition, a gatekeeper capable of routing H.323 calls can assist in balancing decisions between multiple gateways. For example, if a call is routed through a gatekeeper, then the gatekeeper can reroute the call to one of many gateways based on the appropriate routing logic. .. [0270]

Although the gatekeeper is logically independent of the H.323 endpoint, vendors can integrate gatekeeper functionality into the physical implementation of the gateway and MCU. [0271]

Gatekeepers are not required on H.323 systems. However, there is a gatekeeper and the terminal can use the services provided by the gatekeeper. RAS defines these as address translation, admission control, bandwidth control, and zone management. [0272]

The gatekeeper can also serve as a multipoint connection. To support multipoint conferencing, users can employ a gatekeeper to receive H.245 control channels from two terminals to point-to-point conferencing. When the conference is exchanged, or switched, to multipoint, the gatekeeper can redirect the H.245 control channel to the multipoint controller (MC). For this reason, the gatekeeper does not need to process the H.245 signal, it only needs to transfer the signal between terminals or between terminals and MCs. [0273]

A LAN with a gateway can also be equipped with a gateway that translates incoming E.164 addresses into Transport Addresses. Zones are defined by gatekeepers, so an H.323 entity with an internal gatekeeper will have multiple H.323 entities on the LAN with one gatekeeper, and multiple entities will be in the same zone. It may also require a mechanism to disable internal functions so that conditions can be set. [0274]

MSU (Multipoint Control Unit) supports conferencing between three or more endpoints. In H.323, the MCU consists of MC (Multipoint Controller) and MP (Multipoint Processor). MC is required, but MP may not be present or may be multiple. The MC handles H.245 negotiations across all terminals to determine common features for audio and video processing. The MC also controls conferencing resources by determining which audio and video streams to multicast. [0275]

MC does not process media streams directly. This is left to the MP to mix, exchange and process audio, video and / or data bits. The MC and MP functions can exist as dedicated components or as part of other H.323 components. [0276]

The present invention is compatible with RFC1112, 1584, multicast recognition and support for applications such as multimedia, telephone conferencing, databases, distributed computing, real-time workgroups, support for broadcasting capabilities over wireless links, bandwidth. Supports multicast for radio base station 302, including storage, maintaining QoS latency performance, IPv6 IGMP and IPv4 IGMP multicast support, group membership queries, group membership report messages, and more. [0277]

Version 2 of the H.323 standard, approved in January 1998, addresses the shortcomings of version 1, making it a whole new protocol as well as existing protocols such as Q.931, H.245, and H.225. Has also introduced new features. The most important advances are in security, fast call setup, ancillary services, and T.120 / H.323 integration. G. Packet-centric QoS-sensitive wireless point-to-multipoint (PtMP) telecommunications 1. Wireless point-to-multipoint telecommunications system [0278]

Figure 2D shows network 296 with a point-to-multipoint (MtMP) wireless network 298 connected to data network 142 via router 140d. It is important to note that network 296 includes PtMP wireless network 298 in addition to network 286 from Figure 2c. PtMP wireless network 298 will gain access to various audio, data and video resources connected to data network 142 via wireless connection over shared bandwidth through CPE (customer premise equipment) at the subscriber's location. to enable. The wireless PtMP network 298 is a TCP / IP packet-centric (does not generate a dedicated line when transmitting a communication IP flow) packet-switched and QoS-sensitive network. [0279]

In particular, the PtMP wireless network 298 includes a wireless access point (WAP) 290d connected to the router 140d by, for example, a wireless connection. Similarly, the wireless access point 290e can also be connected to the router 140e by wireless connection. The WAP290d wirelessly communicates with one or more radio and transceiver subscriber antennas 292d and 292e, for example by radio frequency (RF) communication. It is obvious to those skilled in the art that various wireless communication methods such as microwave, cellular, spread spectrum, PCS (personal communications system), and satellite communication can be used. [0280]

As an alternative embodiment, RF communication can also be done via cable television (CATV) coaxial cable. It is understandable to those skilled in the art that a coaxial cable functions as a waveguide for propagating RF waves. Therefore, it is possible to make the communication link between the RF transceiver subscriber antenna 292d and the WAP290d a coaxial cable. Since the coaxial connection is similar to the wireless connection, it is given as an alternative form of the wireless connection in the present invention. [0281]

As another alternative embodiment, RF communication can be performed via a satellite connection such as a LEO (low earth orbit) satellite connection or a high orbit satellite. As an example of an LEO satellite connection, the WAP290d and RF transceiver subscriber antenna 292d function as satellite gateways by adding the functionality described in the present invention. [0282]

Although the present invention describes a point-to-multipoint network, it will be obvious to those skilled in the art that the present invention can be similarly applied to a point-to-point network environment. [0283]

FIG. 3A shows one embodiment of the present invention. WAP290d and 290e can be connected to radio base station 302. Radio base station 302 can queue, analyze, characterize, classify, prioritize, and schedule "IP flow" traffic as described later in the drawing. [0284]

FIG. 3B shows one embodiment of the present invention. Antennas 292d and 292e are connected to subscriber CPE (customer premise equipment) stations 294d and 294e, respectively (also called CPE 294d and 294e, respectively). Subscriber CPE stations 294d and 294e are connected to various other CPE devices by wired or wireless connection. For example, CPE stations 294d and 294e are video conferencing devices consisting of voice callers 124d, 124e, 126d, 126e, fax machines 116d, 116e, video monitors 152d, 152e, cameras 154d, and 154e, client computers 120d, 120e, and servers 122d. , 122e can be connected to a host computer. Various legacy devices such as PBXs can also be connected to CPE294d and 294e. In addition, CISCO Systems, Inc., San Next-generation technologies such as Ethernet phones available from Jose, CA's subsidiary Celsius, and other Internet applications can also be connected to CPE294d, 294e via a LAN connection. Other video conferencing devices, such as H.323 compliant conferencing devices, can also be connected to CPE294d, 294e. [0285]

In one embodiment of the invention, either the antenna 292d, 292e can communicate with both the antenna or the WAP 290d, 290e for the backup radio communication path. [0286]

FIG. 3A shows an example of a perspective view 300 of the PtMP network of the present invention. Perspective 300 shows a radio base station 302 that wirelessly communicates with subscriber locations 306a, 306b, 306c, 306d, 306e, 306f, 306g, 306h, 306i, 306j. In particular, the radio base station 302 communicates with the subscriber antennas 292a to 292j at the subscriber locations 306a to 306j via the radio access point 290d. [0287]

The radio base station 302 is connected to the network router 140d by the interface 320 by, for example, a wired connection. The network router 140d is connected to a data network 142 that includes various other network routers 140b that route traffic to other nodes of the data network 142, such as the telephone gateway 288b. [0288]

FIG. 3B shows block diagram 310 showing the wireless PtMP of the present invention. Block Figure 310 shows a radio base station 302 connected to data network 142 through interface 320. In addition, the data network 142 is connected to the router 140d and the telephone gateway 288b connected to the EO104b's Class 5 Central Bureau (CO) switch. The IP phone gateway 288b terminates the telephone traffic at the PSTN facility, for example by converting the packet into a TDM (time domain multiplexed) standard telephone signal. The radio base station 302 communicates with the radio CPE294d at the subscriber's location 306d via the antennas WAP290d and 292d. To those skilled in the art, CPE294d is equipped, for example, with one or more host computers without a telephone device, one or more phones without a host computer, one or more host computers, one or more telephone devices, monitors and cameras. It is self-evident that other conditions can be set to include one or more H.323 functional video conferencing platforms that can include a host computer. [0289]

As shown, the CPE294d is connected to, for example, analog telephone telephone devices 124d, 126d, a host computer, a client 120d and a server 122d. The client 120d and server 122d can be connected to the CPE294d via a LAN connection, such as an Ethernet LAN, or a legacy V.35 device 322d that provides a high-speed data connection. Other internet applications that can connect to the data network can also connect to the CPE294d. 2. Networking Protocol Stack Architecture-Wireless IP Network Access Architecture (WINAAR) [0290]

FIG. 4 shows the wireless IP network access architecture (WINAAR) 400 of the present invention. Architecture 400 shows a version of the TCP / IP protocol stack extended to support packet switching, shared bandwidth, IP-centric with wireless PtMP connections, and QoS. The networking protocol stack consists of physical layer (OSI layer 1) 402, data link layer (OSI layer 2) 404, network layer (OSI layer 7) 406, 408, transport layer (OSI layer 4) 410 and application layer (OSI layer). 7) Described in the OSI (Open Systems Interconnection) 7-tier networking protocol stack consisting of 412. a. Physical layer [0291]

As an example of an example, the physical layer 402 is an integrated line (wASIC) for multiple wireless applications, a commercially available 16QAM / QPSK416ASIC, an IMMUNE (Interference Mitigation and Multipath Negotiation) / RF 418 algorithm that minimizes and / or eliminates harmful interference. ASIC, using FH (frequency hopping) 419 ASIC, which provides dynamic and adaptive multi-channel transmission that optimizes data link integrity by changing the frequency level according to the noise level of a given frequency. Can be executed. Physical layer 402 can include radio frequency (RF) signal 415. b. Data link layer [0292]

The data link layer 404 is located above the physical layer 402. The data link layer 404 can include a MAC layer 414a, a PRIMMA (proactive reservation-based intelligent multi-media access) engineering section 414b, and a MAC (medium access control) layer 414 shown in FIG. 400 as 414c. Arrows 426, 428, and 430 indicate that MAC layer 414 reads header information from data and multimedia application layer 425, TCP / UDP layer 427, and IP layer 429, respectively, and analyzes and schedules "IP flow" IP packets. Shows that can be done. IP packets of an IP flow are identified by analyzing header information to determine the QoS requirements of the IP flow and enable IP flow characterization, classification, representation, prioritization, and scheduling. c. Network layer 1. Internet Protocol (IP) [0293]

Network layer 408 is Internet Protocol (IP) 429. IP is a standard protocol for processing packets of information, as described with reference to data network 142 (more on this later). As shown in FIG. 7, the IP header field 702 can include, for example, source and destination IP addresses, IP TOS (service type, type), IP TLL (expiration date), and protocol fields. IP is a datagram protocol that is very flexible to network failures, but does not guarantee the transmission of sequences. Routers use ICMP (Internet Control Message Protocol) to send error and control messages to other routers. ICMP can also provide the ability for users to send "pings" (echo packets) to check the reachability and round-trip delay of IP address hosts. Another OSI Layer 3 protocol is ARP (address resolution), which can provide an interface directly to the data link layer. protocl). ARP maps a physical address, such as an Ethernet MAC address, to an IP address. 2. Internet Protocol (IP) v4 and v6 [0294]

IP429 of network layer 408 may be, for example, IP version 4 (IPv4) or IP version 6 (IPv6). IPv6 (sometimes called the Next Generation Internet Protocol or IPng) is a backwards compatible and extended version of the current version of the Internet Protocol, IPv4. IPv6 was designed to solve the problems caused by the success of the Internet (lack of address space and router tables). IPv6 also adds the necessary functionality, including line security, automatic conditioning, and real-time services similar to QoS. As the use of the Internet increased and the number of IP addresses assigned increased, it became urgently necessary to increase the address capacity. Since IPv4 uses 32-byte numbers to form addresses, it can provide about 4 billion different network addresses. IPv6, on the other hand, uses 128-byte numbers to form addresses, so it can provide a much larger number of addresses. 3. RSVP (Resource Reservation Protocol) IP429 at network layer 408 can have RSVP extensions. RSVP, which extends IPv4 with QoS features, is supposed to let NetworkManager allocate bandwidth based on the bandwidth requirements of the application. Essentially, RSVP is a new communication protocol that is expected to send bandwidth-reserving signals to routers to send data, video, and audio traffic in real time. [0295]

A resource reservation protocol that operates on a per-connection basis can be used on the network to temporarily raise the priority of a given user. RSVP operates between the ends to communicate application requirements for special processing. RSVP identifies the session between the client and server and requires the router processing the session to prioritize router access for that communication. At the end of a session, the resources reserved for that session are released for others to use. [0296]

RSVP unfortunately can only provide two levels of priority in the signaling scheme. Packets are identified on each router as low priority or high priority. However, if the network is congested, it is not enough to classify it into two levels. In addition, packets prioritized by one router may be rejected by the next router. [0297]

An IETF standard approved in 1997 put RSVP out of control over who receives bandwidth, but the question remains what to do if multiple users request a large amount of bandwidth at the same time. .. Current technology has outlined this situation as being serviced by the first person. The IETF has formed a project team to consider this issue. [0298]

Many people equate QoS with a protocol because RSVP provides a special level of service. For example, Cisoco is currently using RSVP, an IPv4-based internetworking router, to carry IPv6-type QoS features. However, RSVP is only a small part of the QoS picture, as RSVP is only valid to the extent supported within a given client / server connection. RSVP allows applications to demand latency and bandwidth, while RSVP provides traffic flow management with the congestion control or network-scale priorities needed to integrate QoS within an enterprise. I will not do it. Moreover, RSVP does not address the challenges associated with transmitting packets over wireless media. [0299]

The present invention conforms to (1) RFC2205, (2) path message, reservation (Reservation (Resv) message, path teardown message, reservation discard message (resv teadown message), path error. Recognize and support RSVP messages such as path error messages, Resv error messages, and confirmation messages such as confirmation messages, (3) null, session, RSVP_hop, time_values, style, Recognizes and supports RSVP objects such as flowspec, sender_template, sender_Tspec, Adspec, Error_Spec, Policy_data, Integrity, Scope, Resv_Confirm, (4) Provides RSVP Flowspecs conditional conversion for QoS resource allocation for radio base station 302, By supporting RSVP. [0300]

The present invention includes (1) RFC2474, 2475 support, (2) Internet core DiffServ, (3) RSVP / int-serv for host and edge networks, (4) DiffServ compatibility admission control capabilities, (5). ) Differentiated Services (DS) (Field Marking Supported for Use by Differentiated Services, Conversion to Radio Base Station 302 Resource Allocation), (6) Support for Converging Multiple End-to-End Sessions into One Tunnel Session, Provides support for DiffServ and RSVP / int-serv. 4. Real-time Transport Protocol (RTP) and Real-time Control Protocol (RTCP) [0301]

TCP at transport layer 410 can have RTP and RTCP extensions. Real-time Transport Protocol (RTP) is a new protocol endorsed by the IETF Audio / Video Transport Working Group. As shown in Figure 7, the RTP and RTCP header fields 708 can contain multiple information subfields. RTP supports the interactive real-time transmission of audio and video over packet-switched networks. RTP is a small protocol that provides content identification, packet sequencing, timing reconstruction, loss detection and security. RTP allows data to be transmitted to one or more destinations with limited delay. [0302]

RTP and other Internet real-time protocols, such as Internet Stream Protocol Version 2 (ST2), focus on the efficiency of data transmission. RTP and other Internet real-time protocols, such as RTCP, are designed for persistent, high-volume data exchange communication sessions. RTP does not handle resource reservation or QoS control. Instead, RTP relies on resource reservation protocols such as RSVP to communicate and dynamically allocate the appropriate bandwidth. [0303]

RTP adds headers and time stamps that distinguish whether IP packets are data or voice, allowing voice packets to be prioritized. RSVP, on the other hand, allows networking equipment to reserve bandwidth for the transmission of seamless multimedia data streams. [0304]

Real-time Transport Protocol (RTCP) is a protocol similar to RTP that analyzes the state of a network. RTCP operates in a multicast manner, providing feedback to RTP data sources and all session participants. RTCP can be adopted to prevent voice-over-IP datagram transmission from flowing into a dedicated IP network. RTCP notifies applications of spikes or fluctuations in network transmission so that the software can adapt to changes in network load. RTCP network feedback allows telephone software to switch (switch) compression algorithms as the connection deteriorates. 5. IP multicasting protocol [0305]

IP429 at network layer 408 can also support multicasting protocols. Digitized audio and video consist of large amounts of data that, when split into multiple packets, need to be transmitted in the correct order in a timely manner to maintain the quality of the original content. The protocol was developed around providing a way to efficiently send content to multiple recipients. Sending to multiple recipients is called multicasting. Multicasting involves broadcasting a message from one host to many hosts in a one-to-many relationship. The network device broadcasts the message to the PCS or a group of other selected devices such as LAN, WAN, and workstations on the Internet. For example, a router can send information about routing table updates to other routers in the network. [0306]

Multiple protocols have been implemented for IP multicasting, including upgrades to the Internet Protocol itself. For example, the latest version of IP, some of the changed parts of IPv6, are unicast (point-to-point communication), anycast (communication with the closest device group), and multicast. Supports various forms to handle. IP multicast support consists of several protocols such as IGMP (Internet group management protocol), PIM (protocol-independent multi-cast), and DVMRP (distance vector multi-cast routing protocol). Queuing algorithms can be used to ensure that video or other multicast data types arrive on time without visible or audible distortion. [0307]

The Real-Time Transport Protocol (RTP), which is currently in the drafting stage of the IETF, is designed to transmit data such as video and audio end-to-end, in real time. RTP operates by UDP (User Datagram Protocol) and does not guarantee timely transmission, quality of service (QoS), transmission, or transmission sequence. RTP works with mixers and translators to support encryption and security. RTCP (Real Time Transport Protocol) is part of the RTP that defines the analysis of network state. RTCP mandates service monitoring and collects participant information. RTP dynamically communicates with RSVP and allocates the appropriate bandwidth. [0308]

Internet packets usually travel on a first-come, first-served basis (first-come-first-served basis). When the network is congested, RSVP (Resource Reservation Protocol) allows certain types of traffic, such as video conferencing, to be delivered in preference to less time-sensitive traffic, such as email, although it can be expensive. It has become. RSVP may change the pricing structure of the Internet by offering different QoS at different prices. SLA can be used to provide various QoS levels to users on the CPE station side depending on the SLA subscriber level. [0309]

On behalf of the application, the host can use the RSVP protocol to request specific QoS from the network for a specific data stream or flow. Routers can use the RSVP protocol to transmit QoS control requests to all required network nodes to establish and maintain the conditions required to provide the requested services. RSVP requests generally, but not necessarily, reserve resources for each node along the data path. [0310]

RSVP itself is not a routing protocol. RSVP is designed to work with current and future unicast and multicast routing protocols. RSVP processing queries the local routing database and sets the route. For example, in the case of multicast, the host sends an IGMP message to join the multicast group and then sends an RSVP message to reserve resources along the transmission path of that group. The routing protocol determines the destination of the packet. Since packets are forwarded along their routing, RSVP is only concerned with the QoS of those packets. The present invention transmits QoS-sensitive wireless PtMP access to the user over the shared wireless bandwidth and the priority in the packet header of the packet of the IP flow received to transmit over the bandwidth of the wireless base station. Information can be considered. d. Network layer VPN network (example of optional protocol) [0311]

The point-to-point protocols 420 and IPsec 422 of the optional virtual private network (VPN) protocols shown in network layer 406 are described below. [0312]

There are multiple protocol standards for VPN today. For example, there are IPsec (IP security), PPTP (point-to-point tunneling protocol), L2F (layer 2 forwarding protocol), and L2TP (layer 2 tunneling protocol). The IETF has proposed an Internet Protocol (IP) security architecture that can be used to secure Internet-based VPNs, allowing vendors to use many cryptographic algorithms, key lengths, and key escrow technologies. IPsec facilitates the security of private sessions over the Internet between organizational firewalls by encrypting traffic when outgoing to the Internet and decrypting the traffic on the called end. IPsec aims to enable enterprises to mix-and-match the best firewall, encryption, and TCP / IP protocol products. [0313]

IPsec is designed to link two LANs via an encrypted data stream that connects the Internet. 1.PPTP (point-to-point tunneling protocol) [0314]

PPTP (Point-to-point tunneling protocol) provides an alternative approach to VPN security besides the use of IPsec. Unlike IPsec, which is designed to link two LANs through an encrypted data stream that connects the Internet, PPTP allows users to connect to an organization's network over the Internet by a PPTP server or an ISP that supports PPTP. To enable. PPTP was proposed to the IETF as a standard in early 1996. Firewall vendors are expected to support PPTP. [0315]

PPTP was developed by Microsoft in collaboration with 3Com, Ascend, and US Robotics and is now implemented in Windows NT Server 4.0, Windows NT Workstation 4.0, upgraded versions of Windows 95, and Windows 98, Microsoft (Redmond, Washington). It is available from. [0316]

PPTP "tunneling" means encapsulating a message so that it is encrypted and then sent to the Internet. PPTP collaborates on processing resources to create a tunnel between the server and the client. 2.L2F (layer 2 forwarding protocol) [0317]

Developed by Cisco, Layer 2 Forwarding Protocol (L2F) is a PPTP in that it encapsulates other protocols inside TCP / IP packets and transmits them to the Internet or other TCP / IP networks such as Data Network 112. It's similar. Unlike PPTP, L2F requires a special L2F compliant router (which can request changes to the LAN or WAN infrastructure), operates on a low level network protocol stack, and requires TCP / IP routing. It is a point that does not. L2F also adds security to usernames and passwords that PPTP did not have. 3.L2TP (layer 2 tunneling protocol) [0318]

L2TP (layer 2 tunneling protocol) is a combination of L2F specifications and PPTP specifications. In November 1997, the IETF approved the L2TP standard. Cisco has incorporated L2TP into its Internet operating system software, and Microsoft has integrated it into Windows NT 5.0. A major advantage of L2TP over IPsec is that L2TP covers TCP / IP transmission and can carry multiple protocols. L2TP also provides the ability to transmit over non-IP networks. L2TP, however, ignores important security features for data encryption and network administrators in order to adopt a reliable VPN. 4. IPsec [0319]

The present invention supports an IP flow using the security encryption function of IPsec 422. The IPsec 422 flow integrated into the WINAAR architecture 400 is shown in Figures 17A and 17B for the downlink and uplink directions, respectively. The radio base station 302 supports the priority of the IPsec encrypted stream by providing a firewall in the radio base station, decoding the data stream and the packet header information, and then performing identification analysis. The frame stream already contains the encryption of the frame data via the wireless transmission medium and performs frequency hopping. [0320]

IPsec provides secure data transmission for, for example, VPN and eCommerce security. IPsec is compatible with RFC 2401-2407. IPsec is also supported in IPv4, IPv6, and even IPsec tunnnel mode. Radio Base Station 302 Security Protocol Support includes AH (authentication header) and ESP (encapsulating security payload). Radio base station 302 supports IPsec authentication (MD5), encryption algorithms, and automatic key management (IKE and ISAKMP / Oakley). Radio base station 302 selectively provides transport mode and tunnel mode, for example, providing one encrypted tunnel for all traffic between two hosts, or independent encryption for each TCP connection between hosts. You can choose the granularity of security services, such as whether to provide a tunnel. e. Transport layer 1. TCP / IP (transmission control protocol / internet protocol) and UDP / IP (user datagram protocol / internet protocol) [0321]

As mentioned above, Internet Protocol (IP) has become the mainstream networking protocol today. This is largely due to the Internet, which is based on TCP / IP (transmission control protocol / internet protocol), which is a protocol of the same system. TCP / IP is the most common way to connect between PCs, workstations and servers. TCP / IP is part of many software products, including desktop operating systems (eg Microsoft's Windows 95 or Windows NT) and LAN operating systems. [0322]

The most popular LAN protocol today is Novell's NetWare Network Operating System (NOS) IPX / SPX. However, IPX / SPX has been pushed by TCP / IP. Novell now incorporates original IP support into NetWare, eliminating the need for NetWare to encapsulate IPX packets when transmitting over TCP / IP connections. UNIX® and Windows NT servers can also use TCP / IP. Banyan VINES, IBM OS / 2, and other LAN server operating systems can also use TCP / IP. [0323]

The transport layer (4th layer) 410 is TCP (transmission control protocol) or UDP (user datagram), which is a part of the TCP / UDP / IP protocol series suitable for networking protocols. protocol) 427 can be included. As briefly described in Data Network 142, TCP is a standard protocol that segments traffic into packets between source and destination IP addresses and sends, reassembles, and retransmits packets of information. As shown in Figure 7, the TCP header field 706 is, for example, source, destination port number, window size, emergency pointer, flags (SYN, ISN, PSH, RST, FIN), maximum segment size (MSS). ) Can be included. Both TCP and UDP provide the functionality of a TCP / IP host that identifies multiple applications by port number. TCP can provide reliable, continuous data transmission to applications. TCP can also provide adaptive flow control, segmentation, reassembly, and prioritization of data flows. UDP only provides unresponsive datagram functionality. RFC1889, a recently defined real-time protocol (RTP), can provide, for example, real-time capabilities to support multimedia applications. [0324]

TCP uses window-based flow control. Each TCP source dynamically changes the transmission window that determines the number of packets that can be sent per RTT (successive round-trip time). The TCP source can continue to increase the send window if there was no packet loss within the last RTT. When congestion is detected, the source TCP slows down the transmission several times, or "backs-off" it. Increasing the width of the so-called TCP window per hour also increases the burst of packets accordingly. TCP's window flow control protocol has the effect of increasing throughput and buffer utilization until the end of a period of rapid backoff due to packet loss. [0325]

TCP operates over IP and provides reliable end-to-end data transmission over network 142. TCP controls the amount of unresponsive data in motion by dynamically reducing the size of windows or segments. The opposite operation is also performed. If the network element you are communicating with has a low error rate, supports large packets, has sufficient buffers, and supports larger window sizes, increase the size of the window or segment to increase throughput. f. Application layer [0326]

The application layer 412 includes, for example, HTTP (hypertext transport protocol) by TCP, FTP (file transfer protocol), TELNET remote terminal login, SMT (simple mail transfer protocol), and SNMP (simple network management protocol) by UDP, RPC, You can have application 426 including NFS, TFTP. Other applications also include worldwide web browsers such as Netscape Navigator available from AOL (Reston, VA), spreadsheet application programs such as Lotus123 available from IBM (Armonk, NY), and Microsoft (Redmond). , WA) It can run on network stacks such as NetMeeting. Packets sent by those applications may require special processing or prioritization to obtain appropriate end-user QoS. 3. PRIMMA system IP flow prioritization Scheduling mixed IP flows [0327]

FIG. 6 shows a block diagram 600 showing the scheduling of mixed IP flows. Block diagram 600 shows the scheduling of radio base station 302. Block Diagram 600 features include Internet, VPN, and PRIMMA management of real-time IP flows. As shown in FIG. 3A, the radio IP flow is transmitted from the data network 142 to interface 320 of radio base station 302 via network router 140d. The IP flow is then scheduled and transmitted from antenna 290d of radio base station 302 via antenna 292d to subscriber location 306d. [0328]

Block Figure 600 in FIG. 6 shows the downlink and uplink flows between interface 320 and radio base station antenna 290d. As shown here, an IP flow represents a series of data packets sent from a source to a destination post computer. The IP flow 630 transmitted from the data network 142 via (interface 320) is composed of an Internet IP flow 608, a VPN IP flow 610, and a real-time IP flow 612. IP flow 630 is in the downlink direction. [0329]

The downlink IP flow analyzer 602 (hereinafter referred to as "downlink flow analyzer 602") analyzes Internet IP flow 608, VPN IP flow 610, and real-time IP flow 612. The IP flow analyzer 602 will be described with reference to FIGS. 8A and 15A. When the IP flow analyzer 602 receives a packet, it analyzes the packet header field to identify whether it is a new IP flow or an existing IP flow. The IP Flow Analyzer 602 can also characterize the QoS requirements of the IP flow according to the content of the packet header fields. The IP Flow Analyzer 602 can classify IP flows, correlate packets from existing IP flows, and group IP flows with similar QoS requirements. The IP Flow Analyzer 602 can also submit IP flows to the flow scheduler. [0330]

The downlink PRIMMA MAC IP flow scheduler 604 (hereinafter referred to as "downlink flow scheduler 604") schedules the received IP flows 608, 610, and 612 to be transmitted in the downlink direction. The downlink flow scheduler 604 can prioritize different classes of IP flows. For example, the scheduler 604 can reserve a slot for a downlink frame for a latency sensitive IP flow. The scheduler 604 can allocate a large amount of bandwidth to the FTP IP flow 608 for file transmission. E-mail IP flow 608 can be given low priority to packets. When allocating radio band frame slots according to priority, the downlink scheduler 604 uses a VPN from a virtual private network (VPN) where the IP flow 630 is from a remote branch office to a corporate network, for example. It can be considered that the IP flow is 610. All traffic from a VPN can be given a high priority, and certain types of VPN traffic can require a particular level of service. The downlink flow scheduler 604 can prioritize the real-time IP flow 612 so that the real-time IP flow 612 arrives at CPE 294 at CPE subscriber position 306 at the required time. [0331]

The downlink PRIMMA MAC SAR (segmentation and resequencing) and framer 606 (hereinafter referred to as "downlink SAR and framer 606") send the received IP flow data packet to CPE 294 at CPE subscriber location 306 via wireless media. Segment into frames for transmission. For example, IP flows 616, 624 can be transmitted to CPE294d at CPE subscriber location 306d via wireless media via base station 290d (to subscriber antenna 292d and CPE294d at CPE subscriber location 306d). In the present invention, the term wireless medium includes not only propagation of RF transmission by cellular communication, but also RF transmission and cable (eg, coaxial cable) communication widely via satellite communication. [0332]

In the uplink direction, IP flow 626 from CPE294d at CPE subscriber position 306d is received by radio base station antenna 290d. The IP flow 626 can include an internet IP flow 618, a VPN IP flow 620, and a real-time IP flow 622. The uplink IP flow analyzer 632 (hereinafter referred to as the "uplink flow analyzer") analyzes Internet IP flow 618, VPN IP flow 620, and real-time IP flow 622. The uplink flow analyzer 632 will be described with reference to FIGS. 8B and 15B. In one embodiment, the function of IP Flow Analyzer 632 operates at CPE294d at CPE subscriber position 306d and sends data to radio base station 302, including information about the IP flow that CPE294d wants to schedule uplink slots. Send the request. [0333]

The uplink PRIMMA MAC IP flow scheduler 634 (hereinafter referred to as the "uplink flow scheduler 634") can schedule the requested IP flow. In one embodiment, the function of IP Flow Scheduler 634 can be performed at CPE294d at CPE subscriber position 306d. In another embodiment, the function of the IP flow scheduler 634 can be performed on the radio base station 302. The advantage of deploying the uplink flow scheduler 634 in a radio base station is that it is particularly effective in a point-to-multipoint architecture. Placing one centralized scheduler at base station 302 is more effective than deploying multiple uplink flow schedulers 634 at CPE 294 at location 306 for multiple CPE subscribers. [0334]

Uplink PRIMMA MAC SAR (segmentation and resequencing) and framer 636 (hereinafter referred to as "uplink SAR and framer 636") send IP flow data packets via radio medium from CPE 294 at CPE subscriber location 306 to a radio base station. Since it is transmitted to 302 and further transmitted via the data network 142, it is framed. IP flow 626 from CPE294d at CPE subscriber position 306d to base station antenna 290d, subscriber antenna connected to CPE294d at CPE subscriber position 306d via wireless media such as RF communication, cable modem, satellite communication, etc. It can be sent from 292d. b. Overview of downlink and uplink subframe prioritization [0335]

Block Figure 800 in Figure 8A outlines typical downlink analysis, prioritization, and scheduling functions. Similarly, block diagram 830 in FIG. 8B outlines typical uplink analysis, prioritization, and scheduling functions. Block diagrams 800 and 830 illustrate the functions of block diagram 600 in FIG. 6 in more detail. [0336]

First, with reference to block Figure 800 (Figure 8A), prioritization and scheduling of IP flows for shared bandwidth is done from data network 142 via router 140d, interface 320, radio base station 302, WAP290d, radio medium. The wireless transceiver subscriber antenna 292d, the subscriber CPE location 306d, and the subscriber CPE station 294d are described in a downlink path. [0337]

The IP Flow Analyzer 602 identifies, characterizes, classifies, and presents data packets to the downlink frame scheduler. The ability to identify, characterize, classify, and present data packets is shown in Figure 15A. [0338]

When identifying a packet, the field in the packet header section indicates whether the data packet of the incoming IP data flow is known to the system, that is, whether it is an "existing IP flow" or the first data packet of a new IP flow. Judgment based on. This identification process can also include, for example, determining the source of the packet and inferring the information type of the packet payload. [0339]

When characterizing a packet, a new data packet (new IP data flow) that is determined to be unknown to the system is characterized based on the packet header information, the QoS requirements for the IP data flow are determined, and the IP data flow is received. Identify the subscriber CPE station to do. [0340]

When classifying packets, classify the new IP data flow into communication priority classes. Classification can also include grouping different IP flows with similar characteristics into the same class. The class grouping for IP Flow 630 is shown in IP Class 810a-810g. [0341]

When the packet is presented, the new IP data flow is initialized and presented to the downlink flow scheduler 604. [0342]

The downlink flow scheduler places IP data flow data packets on the class queue based on the class queue priority and follows a set of conventions. The pre-booking algorithm schedules data packets to be transmitted over the radio medium to the subscriber CPE station 294d at the subscriber CPE location 306d. Terms include, for example, hierarchical class-based priorities, virtual private network (VPN) directory enabled data priorities (including, for example, DEN (directory enabled networking)), and service level agreements (SLAs). level agreement) Determined based on input to the downlink flow scheduler based on priority. The pre-booking algorithm used for scheduling, for example, isochronous traffic is shown in Figure 14. [0343]

The SAR and Framer 606 split, order, and frame data packets for wireless communication from the WAP290d over the wireless medium to the wireless transceiver subscriber antenna 292. Block As shown in Figure 800, multiple subscriber applications 820a-820e connected to subscriber CPE station 294a-e (not shown) located at subscriber CPE locations 306a-306e, eg, subscribers. Operates on devices such as workstation 120d (not shown). Each subscriber CPE location 306 can accommodate one or more subscriber CPE stations 294, and each subscriber CPE station 294 sends and receives one or more IP data flows to and from one or more subscriber workstations 120. can do. In fact, each application connected to a single CPE station can send and receive multiple IP data flows. [0344]

At position 306 of the subscriber CPE, as shown in FIG. 8a, the CPE SAR and framer 814a sort the received data and send it to the subscriber application 820a via the CPE flow scheduler 816a, CPE IP flow analyzer 818a. The CPE flow scheduler 816a-816e performs the same function as the downlink flow scheduler 604 for uplink traffic. Similarly, the CPE IP Flow Analyzer 818a-818e performs the same functions as the Downlink Flow Analyzer 602. [0345]

In one embodiment of the invention, in downlink mode, the CPE IP Flow Scheduler 816a-816e and CPE IP Flow Analyzer 818a-818e perform no function. [0346]

Block diagram 800 shows the logical functions performed in the downlink path, although not necessarily the physical location of these functions. [0347]

The functions of the subscriber application 820a-820e, CPE SAR and framer 814a-814e can be performed on the actual subscriber CPE station 294 connected to the radio base station 302 via a wireless connection. [0348]

Block Figure 800 shows example 812 of the priorities used by the downlink flow scheduler 604 to place the received data packets on the priority class queue. An example of the following priorities is shown. Latency Sensitive UDP Priority 812a, Priority-High 812b, Priority-Medium 812c, HTTP Initial Screen Priority 812d, Latency Neutral Priority 812e, FTP (file transfer protocol), SMTP (simple mail transfer) protocol) Other email traffic priority 812f, priority-low 812g. One of skill in the art can recognize that many different priority classes are possible depending on the QoS requirements of the end user. Latency-sensitive UDP priority data means data that is sensitive to jitter (jitter: fluctuations caused by synchronization time errors) and latency (latency: transit time required for IP data flow in both directions). Can be done. Priority-High 812b can mean, for example, a premium VPN service, a high priority SLA service, and so on. Priority-Medium 812c can mean, for example, Value VPN service level and Medium level SLA service. HTTP screen priority 812d can mean downloading HTTP data, for example, initial HTTP screen download (important for Internet users to feel that a lot of bandwidth is available for their Internet session). it can. The latency-neutral priority 812e can mean latency-neutral data, such as email traffic. FTP, SMTP priority 812f data contains data that is not sensitive to latency or jitter, but requires a large amount of bandwidth to make an accurate download due to transmission size. Finally, priority-low data 812g can mean data that can be sent over a long period of time, much like one network device sends status information to another network device every 24 hours. [0349]

Block Figure 830 (Figure 8B) runs on the uplink path from subscriber CPE station 294d to radio transceiver subscriber antenna 292d, WAP290d via radio medium, radio base station 302, interface 320, router 140d, data network 140. It shows how to analyze, prioritize, and schedule the IP flow of shared radio bands. [0350]

Block diagram 830 includes an uplink flow analyzer 632, an uplink flow scheduler 634, an uplink SAR and a framer 636. These components are similar to the Downlink Flow Analyzer 602, Downlink Flow Scheduler 604, Downlink SAR and Framer 606, but the data transmitted from the Subscriber Workstation 120 of Subscriber CPE Station 294 over the radio medium. The packet is analyzed, scheduled, framed, and the data packet is sent to interface 320 and sent to data network 142. [0351]

As shown in Figure 8B, the subscriber application 820a-820e is the same as the application shown in Figure 6A. Also, as shown, there are CPE IP Flow Analyzer 819a-819e, CPE IP Flow Scheduler 817a-817e, CPE SAR and Framer 815a-815e. These components function similarly to the subscriber application 820a-820e, CPE IP Flow Analyzer 818a-818e, CPE IP Flow Scheduler 816a-816e, CPE SAR and Framer 814a-814e. However, these features provide an uplink path from the subscriber CPE station (subscriber CPE location 306a-306e) to radio base station 302 to route to the destination host workstation 136 (not shown). Analyze, schedule, and send your IP flow. [0352]

As mentioned above, multiple applications can connect to one or more subscriber CPE stations at subscriber CPE locations 306a-306e. In order to avoid congestion between a plurality of applications competing for a fixed number of bandwidth allocations for uplink communication, one embodiment of the present invention uses a reserved scheduling system. The bandwidth allocation for the data packet is the called frame slot, which will be described later with reference to Figures 12A-12Q, 14, 16A, 16B. [0353]

Block diagram 830 shows the logical functions performed on the uplink path, although not necessarily the physical locations of these functions. [0354]

For example, in one embodiment, the analytical function of the IP Flow Analyzer 632 to identify, characterize, and classify packets for uplink is at subscriber CPE location 306a-306e, subscriber CPE station 294a-294e (illustrated) It is preferred to run on the CPE IP Flow Analyzer 819a-819e in preparation for (not). [0355]

Further, in one embodiment, the function of the CPE IP flow scheduler 817a-817f that schedules the uplink subframe slot is applied to each of the subscriber CPE stations 294 connected to the radio base station 302 via the wireless connection. On the other hand, it can be executed by the radio base station 302. [0356]

In this embodiment, the scheduling function is executed by the uplink flow scheduler 634 of the radio base station 302 based on the classification information provided from the CPE station to the radio base station 302 through the uplink IP flow reservation request. By allocating all scheduling functions to the radio base station 302 and centralizing scheduling control, the overall quality of service system can be optimized. [0357]

However, in other embodiments, each function can be performed at the actual subscriber CPE station. [0358]

In the reservation scheduling function of this embodiment, each subscriber CPE station uses the RRB (Reservation Request Block) of the TDMA air frame, which will be described later with reference to FIG. 12A-12O, of the frame slot for uplink transmission. After requesting a reservation, interface 320 enables communication with an uplink path. After the reservation request, the uplink flow scheduler 634 sends an uplink data packet to the subscriber CPE station 294 using the CPE station 294 from the source subscriber workstation 120, as indicated by the line 640. The description of one or more slots that can be transmitted over the radio medium by the data network 142 in the direction of the destination host workstation 136. c. Service level request [0359]

Figure 9 shows how the PRIMMA MAC IP Flow Scheduler 604 schedules frame slots and allocates resources in consideration of service level agreements. Figure 9 shows Figure 900 of SLA-based IP flow management, including prioritization of uplink traffic sent from CPE subscriber locations 306a, 306b, 306c, 306d to radio base station 302. For example, suppose a telecommunications service subscriber subscribes to one of four SLA levels, P1 902a, P2 904a, P3 906a, and P4 908a. As shown in the example in the figure, the IP flow 902b is sent to the subscriber at CPE subscriber location 306a and has the SLA priority level of P1902a. Similarly, IP flows 902b, 906b, 908b are sent to subscribers at CPE subscriber locations 306b, 306c, 306d, with SLA priority levels of P1904a, 906a, 908a, respectively. PRIMMA MAC schedulers 604 and 634 of radio base station 302 consider SLA-based priorities as subscriber CPE Bandwidth can be allocated to IP flows 902b, 904b, 906b, 908b. In the example shown, IP flow 902b can be assigned to frame slot 902c based on SLA priority 902a. Frame slots 904c, 906c, 908c can also be scheduled with SLA priorities in mind. In this way, uplink IP flow traffic can be sent over the data network 142. [0360]

Priority based on SLAs can provide telecommunications providers with a valuable means of providing differentiated services to a variety of customers. For example, prioritize low-priority traffic from subscribers who have purchased a premium SLA service contract over high-priority traffic from subscribers who have only subscribed to a value-level or low-priced SLA service priority. It is possible to make it possible to schedule. d. Header identification [0361]

As shown in Figure 7, packet header information 700 can be used to distinguish between IP flows and IP flow QoS requirements. In particular, the IP header field 702 is used, for example, to assist the PRIMMA MAC in the source and destination IP address fields, packet or IP flow classifications used by the application to perceive and make preferred resource allocations. It can include IP service type (TOS) fields, IP expiration (TTL) fields used to predict application packet drops, and protocol fields that can be used to identify IP flows. [0362]

Packet header information 700 also includes UDP header field 704. The UDP packet header field contains the source and destination port numbers. [0363]

Packet header information 700 also includes TCP header field 706. The TCP packet header field 706 includes source and destination port numbers, TCP sliding window size, emergency pointer, SYN, ISN, PSH, RST, FIN flag, and maximum segment size (MSS). [0364]

Packet header information 700 can also include real-time protocol RTP and RTCP header field 708. [0365]

It is self-evident to those skilled in the art that other packet header fields that identify the IP flow are available. These fields are given as an example and are not an exhaustive list of available packet header fields. Other fields, such as fields from IPv6 related to DIFF SERV (Differentiated Services), can also be used with the IP flow analyzers 602, 632 of radio base station 302. e.TDMA MAC air frame [0366]

Figure 12A-12O shows an example of a TDMA (domain multiple access) MAC (media access control) transmit airframe. The fields illustrated herein merely indicate one embodiment of the present invention and are not intended to limit the present invention. [0367]

Figure 12A shows the entire TDMA MAC transmit airframe. The airframe 1202 includes a downstream transmission subframe 1202 and an upstream transmission subframe 1204. [0368]

The TDMA MAC transmit airframes in Figure 12A are UAB (Upstream Acknowledgment Block) 1206, ARB (Acknowledgment Request Block) 1208, FDB (Frame Identifier Block) 1210, Data Slot (DS).₁1212a, DS₂1212b, DS₃1212c, DS₄1212d, DS₅1212e, DS₆1212f, DS₇1212g, DS₈1212h, DS₉1212i, DS₁₀1212j, DS₁₁1212k, DS_m1212l, DAB (Downstream Acknowledgment Block) 1214, RRB (Reservation Request Block) 1216, UA₁1218a, UA₂1218b, UA₃1218c, UA₄1218U, UA₅1218e, UA₆1218f, UA₇1218g, UA₈1218h, UA₉1218i, UA₁₀1218j, UA₁₁1218k, UA₁₂1218l, UA_mIncluding 1218m. [0369]

The examples described here are TDMA types used in TDMA / Time Division Duplex (TDMA / TDD). In TDMA / TDD, transmission is performed from CPE station 294 to radio base station 302 during a certain time period, and from radio base station 302 to CPE station 194 during another time instance (period). Multiple slots can be used for uplink and downlink. The number of slots assigned to the uplink and downlink can change dynamically. However, the downlink data rate is usually faster than the uplink data rate, so more slots are allocated for the downlink. The slots distributed between the downlink and the uplink are dynamically assigned, but in this embodiment the total number of slots per frame is fixed.

[Table 5]

<img file="JP2003521138A_D0010.tif" /><img file="JP2003521138A_D0011.tif" /> 【0370】

FIG. 12B is a diagram illustrating a typical TDMA / TDD airframe 1220 of the present invention. The frame size 1228 of the TDMA / TDD airframe structure 1220 may be, for example, 16 or 32 slots. It is obvious to those skilled in the art that the number of slots in the frame structure 1220 may be any other number without departing from the spirit and scope of the present invention. The frame structure 1220 includes, for example, various TDMA slots 122a, 122b, 122c, 122d. Each TDMA slot 122a-c can include data slots 1224a, 1224b, 1224c, 1224d, which can sequentially have control packets 1226a or data packets 1226b-d, respectively. [0371]

In this embodiment, the total of TDMA slots 1222 in one frame of frame size 1228 is fixed. However, as described above, using the resource allocation method of the present invention, all uplink TDMA slots are combined into one uplink subframe or upstream transmission subframe 1204 with the total number of TDMA slots 1222 as one subset. Dynamically allocate in the uplink direction, and downlink all downlink TDMA slots as one downlink subframe or downlink transmit subframe 1202, with the total number of TDMA slots 1222 as one subset. It can be dynamically assigned in the link direction. Using the resource allocation method of the present invention, all TDMA slots 1222 can be allocated in a predetermined upstream or downstream direction. In addition, all data slots 1224 can be assigned to a single CPE station. Radio base station 302 is a state machine It has a machine) and knows the state of each connected CPE station 294 (IP flow is recognized by the radio base station 294). [0372]

The downstream transmission subframe 1202 and the upstream transmission subframe 1204 will be described in detail below. 1. Downstream transmission subframe [0373]

Figure 12C shows a typical downstream transmission subframe 1202. The downstream transmit subframe in Figure 12C includes transmitter response times 1230, UAB1206, ARB1208, FDB1210, a different number of DS1212s (eg 16) per frame, and CCB (command and control block) 1232. DS transmission 1212 is DS₁1212a, DS₂1212b, DS₃1212c, DS₄1212d, DS₅1212e, DS₆1212f, DS₇1212g, DS₈1212h, DS₉1212i, DS₁₀1212j, DS₁₁1212k, DS_mIncludes 1212l. [0374]

FIG. 12D is a diagram showing a typical UAB1206 of the downstream transmission subframe 1202. The downstream transmission subframes in Figure 12D are UAB1206, ARB1208, FDB1210, DS.₁1212a, DS₂1212b, DS₃1212c, DS₄1212d, DS₅1212e, DS₆1212f, DS₇1212g, DS₈1212h, DS₉1212i, DS₁₀1212j, DS₁₁1212k, DS_mIncludes 1212l, CCB1232. [0375]

UAB1206 is a subslot UAB₁1206a, UAB₂1206b, UAB₃1206c, UAB₄1206d, UAB₅1206e, UAB₆1206f, UAB₇1206g, UAB_nIncluding 1206h. UAB₁1206a includes a preamble 1234a, a subscriber ID 1234b, an IP flow identifier 1234c, a slot sequence number 1234d, and a CRC (cyclical redundancy check) 1234e. [0376]

The UAB field is a slot for upstream transmission subframes from radio base station 302 to CPE station 294 (eg US).₁-US₁₆) Is received. The upstream transmission subframe will be described. [0377]

ARB1206 subslot UAB₁In 1206a, the preamble 1234a contains the data used for link integrity, the subscriber ID 1234b identifies the CPE station 294 making the reservation request, the IP flow identifier 1234c identifies the IP data flow, and the quality of service. Data class 1234a identifies the priority class of the IP data flow if known to CPE station 294, IP flow priority and type 1234b indicate that it is a new IP data flow, and CRC1234e, which means cyclic redundancy code, Subslot RRB₁Provides error checking bits for 1216a. [0378]

Figure 12E shows a typical ARB1208 of downstream transmit subframe 1202. The downstream transmission subframes in Figure 12E are UAB1206, ARB1208, FDB1210, DS.₁1212a, DS₂1212b, DS₃1212c, DS₄1212d, DS₅1212e, DS₆1212f, DS₇1212g, DS₈1212h, DS₉1212i, DS₁₀1212j, DS₁₁1212k, DS_mIncludes 1212l, CCB1232. [0379]

ARB1208 is a subslot ARB₁1208a, ARB₂1208b, ARB₃1208c, ARB₄1208d, ARB₅1208e, ARB₆1208f, ARB₇1208g, ARB_nIncluding 1208h. ARB₁1208a includes a preamble 1234a, a subscriber ID 1234b, an IP flow identifier 1234c, a slot sequence number 1234d, and a CRC (cyclical redundancy check) 1234e. [0380]

The ARB field is an acknowledgment that the radio base station 302 has received an upstream reservation request from the CPE station 294 to the CPE station 294. The upstream transmission subframe will be described. [0381]

ARB1208 subslot ARB₁In 1208a, the preamble 1234a contains the data used for link integrity, the subscriber ID 1234b identifies the CPE station 294 making the reservation request, the IP flow identifier 1234c identifies the IP data flow, and the quality of service. Data class 1234a identifies the priority class of the IP data flow if known to CPE station 294, IP flow priority and type 1234b indicate that it is a new IP data flow, and CRC1234e, which means cyclic redundancy code, Subslot RRB₁Provides error checking bits for 1216a. [0382]

FIG. 12F is a diagram showing a typical FDB 1210 of the downstream transmission subframe 1202. The downstream transmission subframes in Figure 12F are UAB1206, ARB1208, FDB1210, DS.₁1212a, DS₂1212b, DS₃1212c, DS₄1212d, DS₅1212e, DS₆1212f, DS₇1212g, DS₈1212h, DS₉1212i, DS₁₀1212j, DS₁₁1212k, DS_mIncludes 1212l, CCB1232. [0383]

FDB1210 is a slot for downstream transmit subframes (eg DS₂-DS₁₆) Contains detailed information. [0384]

The FDB1210 includes a preamble subslot 1236a, a number of downstream slot subslots 1236b, an IP flow ID 1236c for one upstream reserved subslot, and a competing slot count 1236f for the next upstream subframe subslot. [0385]

In FDB1210, the fields are defined as follows. The preamble subslot 1236a contains data used for link integrity, the number of downstream slot subslots 1236b contains the number of downstream slots (DS), and the IP flow ID 1236c for downstream reserved slots is DS.₁IP Flow ID 1236d for Downstream Subslot Includes IP Flow Identification for DS₂IP flow ID 1236e for downstream reserved n subslots, including second IP flow identification for DS_mThe competing slot count 1236f for the next upstream subframe subslot, including other IP flow identification for, provides a count of the next available upstream subframes. [0386]

Figure 12G shows a typical downstream MAC payload data unit (PDU). The downstream MAC PDU contains information about the structure of the current payload. The downstream MAC PDU in Figure 12G includes MAC link list number 1238a (MAC link list number), reservation request index number 1238b (index to downstream IP flow), compressed IP flow identifier 1238c, and compressed IP flow. Includes priority and type 1238d (identifies the priority and type of compressed IP flow), slot payload 1238e (amount of data in downstream data slots), CRC1234e (error checking information). [0387]

Figure 12H shows a typical CCB of downstream transmission subframe 1202. The CCB consists of OAM &amp; P commands ordered by subscriber CPE station 294 for each frame and frame synchronization. CCB1232 has a mode command subslot 1240a (optionally includes the mode taken by the CPE station), a profile command subslot 1240b (including certain system commands such as module patches), and a control data index subslot 1240c (with download location). Memory requirements, or other information required by the CPE station to download data), data block 1 subslot 1240d (including specific system data), data block 2 subslot 1240e (similar), data block n Includes subslot 1240f (similar) and CRC subslot 1234e (error checking information). 2. Upstream transmission subframe [0388]

Figure 12I shows a typical upstream transmission subframe 1204. The upstream transmit subframe in Figure 12I is an ODB consisting of transmitter response time 1230, DAB1214, RRB1216, a different number per frame, eg 16 US1218s, OAM &amp; P data from the subscriber, ordered by the subscriber per frame. (Operation data block) 1242 is included. US transmission 1218 is US₁1218a, US₂1218b, US₃1218c, US₄1218d, US₅1218e, US₆1218f, US₇1218g, US₈1218h, US₉1218i, US₁₀1218j, US₁₁1218k, US₁₂1218l, US_nIncluding 1218m. [0389]

Figure 12K shows a typical RRB1216 of the uplink transmission subframe 1205. The uplink transmission subframes in Figure 12K are also DAB1214, RRB1216, US₁1218a, US₂1218b, US₃1218c, US₄1218d, US₅1218e, US₆1218f, US₇1218g, US₈1218h, US₉1218i, US₁₀1218j, US₁₁1218k, US₁₂1218l, US_n1218m, including ODB1242. [0390]

RRB1216 is a subslot RRB₁1216a, RRB₂1216b, RRB₃1216c, RRB₄1216d, RRB₅1216e, RRB₆1216f, RRB₇1216g, RRB_nIncluding 1216h. RRB₁1216a includes preamble 1234a, subscriber ID 1234b, IP flow identifier 1234c, quality of service (QoS) data class 1244a, IP flow priority and type 1244b, CRC1234e. [0391]

CPE station 294 is a subslot of RRB1216 (RRB₁1216a, RRB₂1216b, RRB₃1216c, RRB₄1216d, RRB₅1216e, RRB₆1216f, RRB₇1216g, RRB_nUse one of 1216h) to make a reservation request, i.e. a request for the bandwidth of future uplink transmission subframes by CPE station 294. When two CPE stations 294d, 294e try to access the same subslot of RRB1216 (which can happen if the pseudo-random number generator chooses the same subslot), "congestion" occurs and the data becomes unreadable at radio base station 302. .. The two CPE stations 294d and 294e will need to be retried. [0392]

A reservation request slot can also be provided for each IP flow. Allow a default number of slots (eg 5) to be used as competing slots instead of assigning one reservation request slot to each CPE subscriber station. If congestion is detected by more subscriber requests than the number of reserved request slots, the allocated slots can be dynamically changed to provide additional RRB slots. (Congestion is similar to Ethernet CSMA / CD congestion, where devices colliding on Ethernet try again at random times and retransmit via the bus architecture.) [0393]

The wireless communication method of the present invention is an improvement on the "Aloha" method developed by N. Abramson in the early 1970s. The so-called "Slotted Aloha" method developed by Roberts in 1972 is based on the concept of the so-called "Aloha" method. It is a bit-mapped reservation protocol. Similar to the Slotted Aloha method, the present invention provides discrete slots for data transmission rather than at any point. However, the present invention has the advantage of transmitting only a "reservation request" that describes the content of the actual data payload, instead of transmitting the actual "payload" of the data. Another advantage is that the number of slots for reservation requests can be dynamically changed according to the latest frequency of detected congestion. [0394]

Unlike various CSMA (Carrier Sense Multiple Access) technologies previously used wirelessly, such as persistent and non-persistent, the method of the present invention allows the subscriber CPE station 294d to carry the carrier (radio channel) before transmitting. Has the advantage of not having to "detect". Instead, the subscriber CPE station 294d selects a "subslot" and selects a pseudo-random number to transmit without carrier detection. If congestion is detected, subscriber CPE station 294d will try again at the next frame using pseudo-random processing. [0395]

Instead of using a bitmap protocol for conflict resolution, as used by some reservation protocols, radio base stations can explicitly approve reservation requests. The standard bitmap protocol requires that all stations be able to receive signals from all other stations so that the order of subsequent transmissions can be indirectly determined from the bitmap pattern. The method of the present invention has an advantage that it is not necessary to receive a reservation request signal from another CPE subscriber station 294d. This is because, for example, the requirement to receive transmissions from other CPE subscriber stations 294d at high frequencies (eg 2GHz to 30GHz) where line-of-sight distance may be limited severely constrains the CPE subscriber station's topology, location, and distance. It is advantageous because it can lead to this. [0396]

Other factors can also be considered, such as the relative or dynamic CPE station 294d (or IP flow) priority factor, because radio base station 302 is now able to explicitly approve the requested reservation. Therefore, by using the reservation protocol of the present invention in which the number of competing subslots can be dynamically adjusted and the explicit reservation approval of the radio base station, radio allocation such as radio and bandwidth is performed according to the QoS request of the IP flow. It is possible to provide a more optimal method than the conventional method. [0397]

As mentioned above, RRB₁1216a includes the following fields: That is, it includes a preamble 1234a, an IP flow identifier 1234c, a quality of service (QoS) data class 1244a, an IP flow priority and type 1244b, and a CRC1234e. RRB1216 subslot RRB₁In 1216a, preamble 1234a contains data used for link integrity, subscriber ID 1234b identifies the CPE station 294 making the reservation request, IP flow identifier 1234c identifies the IP data flow, and quality of service. Data class 1234a identifies the priority class of the IP data flow if known to CPE station 294, IP flow priority and type 1234b indicate that it is a new IP data flow, and CRC1234e, which means cyclic redundancy code, Subslot RRB₁Erachi of 1216a provides a Ekkubitto. Optional subslot RRB₁You can also add to 1216a a field that contains the number of data packets that CPE station 294 sends in its IP data flow. [0398]

FIG. 12J is a diagram showing a typical DAB1214 of the upstream transmission subframe 1204 in which the CPE makes an acknowledgment of slot reception from the base station. DAB is an acknowledgment from subscriber CPE station 294 to the radio base station that a downstream slot was received in the previous subframe. [0399]

DAB1214 is DAB₁1214a, DAB₂1214b, DAB₃1214c, DAB₄1214d, DAB₅1214e, DAB₆1214f, DAB₇1214g, DAB_nIncluding 1214h. Subslot DAB₁1214a includes a preamble 1234a, a subscriber ID 1234b, an IP flow identifier 1234c, a slot sequence number 1234d, and a CRC (cyclical redundancy check) 1234e. (These fields have the same information as described in RRB.) [0400]

Figure 12L shows a typical MAC PDU upstream slot. The MAC PDU upstream slot in Figure 12L has CPE linked list number 1246, CPE linked list number 1246, reservation request index number 1238b, compressed IP flow identifier 1238c, compressed IP flow priority and type 1238d, slot payload 1238e. , Includes CRC1234e. Upstream MAC PDUs are similar to downstream MAC PDUs, but are used for upstream payload information. [0401]

Figures 12M, 12N, and 12O are diagrams illustrating the details of a typical ODM1242. This field can be used to store information about the connection between radio base station 302 and CPE station 294. ODB1242 includes preamble 1234a (including link integrity data), subscriber ID 1234b (identifies the CPE station 294 making the reservation request), system state 1248a (information about the status of CPE station 294), and performance data 1248b (buffer). Includes statistics, cpe processor performance statistics, system state), antenna data 1248c (information about hermitage difficulties), CRC1234c (error check information), synchronization pattern 1248d (error check information). [0402]

As shown in Figure 12M, the system state subslot 1248a has system mode 1250a (CPE station mode, such as command mode, operating mode, system initial mode, etc.), system status 1250b (CPE station status), and system resources. It consists of 1250c (CPE station mode), system power 1250b (CPE station mode), and system temperature 1250a (CPE station temperature). CPE station 294 needs to send information in sequence using ODB1242. [0403]

As shown in FIG. 12N, the performance data 1248a consists of the number of repeated communication attempts 1252a, the number of slipped frames 1252b, and the wait state index 1252c (index to the wait state). f. Frame prioritization based on representative classes [0404]

FIG. 13 is a block diagram 1300 showing a typical flow scheduler for scheduling the present invention. Block Diagram 1300 shows Flow Scheduler 604, 634 (combination of Downlink Flow Scheduler 604 and Uplink Flow Scheduler 634), Downlink Transmit Subframe 1202 (ie, Next MAC Downstream Subframe), Uplink Transmit Subframe. Contains 1204 (ie, the current MAC upstream subframe). Block Figure 1300 also shows the following downstream components: That is, it contains a downstream reserved FIFO (first-in-first-out) queue 1322, a class-1 downstream queue 1302, a class-2 downstream queue 1304, and a class-3 downstream queue 1306. Block diagram 1300 further includes the following upstream reservation components: That is, the current MAC upstream subframe 1344 (current upstream subframe 1204 to save), the previous MAC upstream subframe 1346, 1348, 1350, class 1 upstream reservation request queue 1308, class 2 upstream reservation request. Queue 1310, Class 3 upstream reservation request queue 1312. [0405]

On the downlink path, the IP Flow QoS Class Queuing Processor (discussed later with reference to Figures 15A and 15B) puts the received data packets on Class 1 Packet Flow Queue 1324, 1326, 1328, Class 2 Packet Flow Queue 1330. , 1332, 1334, Class 3 Packet Flow Queues to Queues 1336, 1338, 1340, 1342. [0406]

From Priority Processors Based on Hierarchical Classes, VPN (Virtual Private Network) DEN (Directory enabled networking) Data Tables, SLA (Service Level Contract) Priority Data Tables (discussed later with reference to Figures 15A and 15B) Based on the input in, class 1, class 2, and class 3 packet flow queues are assigned to class 1 downstream queue 1302, class 2 downstream queue 1304, and class 3 downstream queue 1306, respectively. The flow schedulers 604 and 634 schedule these downlink data packets to the downlink transmission subframe 1202. [0407]

In one embodiment, additional processing can be performed to minimize latency and jitter. For example, suppose a data packet in class 1 packet flow queue 1324 requires transmission without jitter or latency, that is, real-time packet transmission at regular time intervals. Packet flow queue 1324 creates four equal slot reservations with equal time intervals in future frames, for example as shown in class 1 downstream queue 1302 (see Figure 14). The reservation is sent to the downstream reservation FIFO queue 1322 and scheduled by the flow schedulers 604 and 634 to future downstream frame 1202. [0408]

In the uplink path, the reservation request for the future upstream slot reaches radio base station 302 via radio medium as part of the current upstream subframe 1204 received from CPE subscriber station 294. The current upstream subframe 1344 can temporarily store a reservation request for analysis and scheduling of uplink packets as described in FIG. 8B. The previous upstream subframes 1346, 1348, 1350 include upstream reservation requests waiting for upstream frame slot allocation in future upstream subframe 1204. The RRB (Reservation Request Block) described with reference to Figure 12 *** contains the reservation of multiple slots for one IP flow of IP identifier # and class of flow. Upstream Reservation Requests (depending on IP flow and class) are in Class 1 Upstream Reservation Request Queue 1308, Class 2 Upstream Reservation Request Queue 1310, Class 3 Upstream Reservation Request Queue 1312 to IP Flow QoS Class Queuing Processor (with Figure 16A). Queued by (see 16B later). The flow schedulers 604, 1566, 634, and 1666 use these downstream and upstream reservation requests to allocate slots to the next downstream transmit subframe 1202 and upstream transmit subframe 1204 data packets, respectively. [0409]

FIG. 14 is a typical two-dimensional block diagram 1400 of the pre-booking algorithm. Figure 14 shows the MAC subframe schedulers 1566, 1666 and frames of data packets sent at time n, n + 1, n + 2, ... n + x, current frame n1402, future frame n +. 1 1404, future frame n + 2 1406, future frame n + 3 1408, future frame n + 4 1410, future frame n + 5 1412, future frame n + 6 1414, ... future frame n Includes + x 1416. Each frame is divided into a variable length downlink subframe 1202 and a variable length uplink subframe 1204. The downlink subframe 1202 and the uplink subframe 1204 together make up the length of the entire frame. [0410]

Each frame n1402 contains multiple slots (1418-1478). Slots 1418-1446 constitute the downlink subframe 1202, and slots 1448-1478 constitute the uplink subframe 1204. In one embodiment, the length of the slots is fixed and one slot can store one data packet. The total number of frame slots in the frame is constant. For example, if a frame has 64 frame slots, those slots are dynamic, for example, 32 to uplink, 32 to downlink, or 64 to uplink, 0 to downlink. Can be assigned in the direction of the uplink or downlink. In block diagram 1400, each slot has one time value (ie, time interval per slot), eg 0.01 ms, and each frame has a frame interval time value (ie, time interval per frame), eg 0.5 ms. You can think of it as a dimensional matrix. [0411]

In the present invention, the pre-booking algorithm allocates future slots to data packets based on the priority of the IP data flow to which the packet is associated. Typical priorities are as described with reference to FIGS. 8A and 8B. Jitter-sensitive calls are time-sensitive (time-sensitive), so it is important to maintain an isochronous connection (in-phase with time). For such signals, it is important that the data is distributed between frames in the same slot, or in slots with periodic variations between frames. For example, vertical reservation 1480 represents a jitter-sensitive signal that receives the same slot for downlink communication at each frame. In particular, the signal is assigned to slot 1422 in frames 1402-1416. If the interval between frames is 0.5ms, this IP flow will be allocated to one slot at 0.5ms intervals. In another example, diagonal reservation 1482 represents a jitter-sensitive signal that receives slots that change over a period of time between consecutive frames. In particular, the signals are assigned to slot 1440 of frame 1402, to slot 1438 of frame 1404, ... to slot 1426 of frame 1416, forming a "diagonal". If the interval between frames is 0.5ms and the interval between slots is 0.01ms, this IP flow will be 0.5-0.01 in one slot and 0. It can be assigned at intervals of 49ms. Therefore, when shortening the frame interval, it is possible to use an oblique reservation that rises to the right. If you want to increase the frame spacing, you can use a downward-sloping diagonal reservation, for example, a downward-sloping diagonal uplink reservation 1486. Diagonal reservation 1482 can be made more prominent (eg, tilted up or down) depending on the setting of the period between consecutive frames. Reservation patterns 1480, 1482, 1484, and 1486 are useful patterns for jitter-sensitive communication. Also, the vertical reservation 1486, similar to the vertical reservation 1480 shown in the figure, is useful for jitter-sensitive communication in the uplink direction. [0412]

In the case of latency-sensitive (latency-sensitive), one or more slots can be guaranteed for each frame. For example, for a call that is sensitive to latency but not to jitter, one (or more) slots can be assigned to each frame for communication. However, the slots do not have to be periodic between frames as in the case of jitter-sensitive calls. The more slots you allocate per frame for an IP flow, the greater the total bandwidth for each frame rate of that IP flow. [0413]

For calls that are less sensitive to latency, fewer slots are allocated per frame for communication. For example, for communications that are less latency sensitive, one slot can receive guaranteed bandwidth for every four frames. For calls that are even less sensitive to latency, you can receive one slot, for example every 10 frames. [0414]

Using these principles, the pre-booking algorithm can use the available number of future frame slots to allocate slots from the highest priority to the lowest priority. Initially, both jitter and latency sensitive IP data flows can be assigned to slots with periodic patterns (eg 1480, 1482, 1484, 1486). Then assign the flows that are very sensitive to latency (but not sensitive to jitter), and finally the flows that are least sensitive to latency in this order. Prioritization of IP flows of different classes by schedulers 604, 634, 1566, and 1666 will be described later with reference to Figures 15A, 15B, 16A, and 16B. g. Downlink subframe prioritization 1. Overview [0415]

Figures 15A and 15B are typical logical flow diagrams for analysis and scheduling of shared radio bandwidth in the downlink direction. The logical flow relates to an IP packet flow arriving from the data network 140 to the radio base station 302 for transmission to the subscriber CPE station 294d via the radio medium. FIG. 15A is a typical logic flow diagram 1500 of the downlink IP analyzer 602. FIG. 15B is a typical logical flow diagram 1560 of the downlink flow scheduler 604. The functional components of FIGS. 15A and 15B will be described by the method module. Method modules can be thought of as physical units (software, hardware, or combinations thereof) or logical means (used for explanatory purposes only). It is self-evident to those skilled in the art that these modules are used only for purposes of explaining this embodiment and are not limiting. [0416]

A typical logical flow of a downlink IP flow analyzer of FIG. 15A Figure 1500 includes a packet header identification unit 1502, a packet characterization unit 1504, a packet classification unit 1506, and an IP flow presentation unit 1508. The details of the functions of each part will be described. [0417]

In one embodiment, the downlink IP flow analyzer 602 is physically located at location of radio base station 302, but those skilled in the art may deploy the same function at a location away from radio base station 302. It is self-evident that it is possible. In a preferred embodiment of the invention, the function of the IP flow analyzer 632 is performed at the subscriber CPE station 24d, which desires an uplink reserved slot for uplinking the packet / IP flow to base station 203. Request block (RRB) requests, including IP flow identifiers, number of packets, and IP flow classification details, can be generated by the IP flow analyzer 632, uplinked by competing RRBs, and wireless by the uplink frame scheduler 634. It is preferred that the base station 302 be scheduled to a future uplink frame. [0418]

Figures 2D, 3A, and 3B are intended to help you understand the downlink IP flow analyzer. 2. Introduction [0419]

The IP Flow Analyzer 602 identifies, characterizes, classifies, and presents data packets to the downlink frame scheduler 604. The functions for identifying, characterizing, classifying, and presenting data packets are executed by the packet header identification unit 1502, packet characterization unit 1504, packet classification unit 1506, and IP flow presentation unit 1508 of the downlink IP flow analyzer 602, respectively. To. [0420]

The packet header identification unit 1502 determines whether the data packet of the incoming IP data flow is part of an IP flow known to the system or is the first data packet of a new IP data flow in the contents of the packet header field. Judge based on. The packet header identification unit 1502 also identifies, for example, the source of the packet from the contents of the packet header field. The packet characterization unit 1504 characterizes a new data packet (or new data flow), determines the QoS requirements for that IP data flow, and identifies the subscriber CPE stations involved in the subscriber workstation receiving the IP data flow. .. The packet classification unit 1506 classifies new IP data flows into communication priority classes and groups packets by IP flows of similar types. IP data flow presentation 1508 initializes a new IP data flow and presents it to the downlink flow scheduler 604. [0421]

The downlink flow scheduler 604 arranges the data packets of the IP data flow in the class queue, and schedules the data packets to be transmitted to the subscriber CPE station via the wireless medium, for example, by a pre-booking algorithm according to a series of conventions. The convention is for input to the downlink flow scheduler from priority module 1574, VPN (virtual private network) DEN (directory enabled) data table 1572, service level agreement (SLA) priority data table 1570 based on hierarchical classes. It can be decided based on. The advance reservation algorithm has been described with reference to FIG. 3. Identification [0422]

The packet header identification unit 1502 identifies the IP flow received from the data network 142 by the data interface 302 based on the packet header. [0423]

An IP flow packet stream from data network 142, including packets from various IP flows (each IP flow is associated with one data "call"), is received by packet header identifier 1502. The IP flow can include packetized data including all kinds of digital information such as packetized voice, video, audio, data, IP flow, VPN flow, real-time flow and the like. The IP flow is transmitted via the data network 142, for example, from the host workstation 136d and reaches interface 302 of the radio base station 320. Interface 302 sends a packet of IP flow to packet header identification unit 1502. On module 1510, received packets are buffered in the storage area. Module 1520 extracts and parses the contents of the packet header fields. [0424]

For IP flows known to the system, so-called "existing IP flows", there is an entry in Table 1526. If the characterized IP data call already exists, the IP flow is in the system. Module 1522 determines if there is a match between the incoming packet and the existing IP flow call of the entry in the existing IP flow identification table 1526. If it is determined to match, the IP flow is known to the system and is sent to module 1530 of packet characterization unit 1504. [0425]

If it is determined that there is no match, the IP flow is a new IP data flow and is sent to module 1524 for analysis of the packet header fields. Module 1524 analyzes the source field of the packet header and determines from the source application's packet header data table 1528 the type of source application that is making the data call or sending the IP packet. .. The application may be the application shown in FIG. 2D or one known to those of skill in the art. As an example, FTP (file transfer protocol), IP voice downloaded from another client workstation 138f Telephony calls (via telephone gateway 288b), voice telephone calls from caller 124d (connected via modem), email from host workstation 136a connected to LAN128a, fax calls, multiple callers 124d , 126d (connected via modem), etc. If the IP flow is not known to the system, the IP flow is given an IP flow identification number and sent to module 1526 to add the IP flow identification number to the existing IP flow identification table 1526. [0426]

When the type of the source application is determined by packet header information or other means such as directly identifying the application, it is sent from module 1524 to module 1532 of packet characterization unit 1504. It is also possible to analyze the TOS (Service Type) or DiffServ (differential service) fields to identify the type of application that is the source of the IP flow. 4. characterization [0427]

Packet characterization unit 1504 characterizes a new IP flow before sending it to packet categorization unit 1506 for classification. [0428]

For existing IP flows, the packet header identification unit 1502 proceeds from module 1522 to module 1530. If module 1522 determines that the IP data flow is known to the system, module 1530 determines if the packet is too old. This can be done, for example, by determining the packet existence time from the time to live packet header field (IP packet header field) and comparing the field with the threshold lifetime value. If the packet is determined to be too old, it can be discarded. Client application discards can be predicted based on packet lifetime. If not, proceed to module 1540 of packet classification unit 1506. [0429]

For the new IP flow, the packet header identification unit 1502 proceeds from module 1524 to module 1532. If Module 1524 determines that the IP flow is unknown to the system, Module 1532 determines the QoS requirements for that application from the source application information identified in Modules 1524 and 1528. Module 1532 performs this action by discovering the QoS requirements of the source application identified with reference to the QoS requirements table 1534. Different applications have different QoS requirements to provide the convenience recognized by the end user. For example, bandwidth allocation (ie, allocating the right amount of bandwidth) is important for applications that perform FTP file transfer downloads, but jitter (time synchronization of received data), latency (between responses). Elapsed time) is not important. On the contrary, jitter and latency are important for voice calls and conference calls, but bandwidth allocation is not. [0430]

After processing by module 1532, module 1536 performs a lookup of the destination CPE subscriber station ID from the subscriber CPE IP address table 1538 for that IP flow. Each subscriber CPE station 294d may have one or more applications running on one or more subscriber workstations 120d homeed therein. Therefore, the IP flow can accommodate one or more applications of one or more subscriber workstations of one or more CPE stations 294d. The subscriber workstation can use the device connected to the subscriber CPE station 294d. Module 1536 refers to the IP flow in Table 1538 to determine the identification of subscriber CPE station 294d receiving packets for the new IP flow from data network 142. After that, the process proceeds from module 1536 to module 1542 of the packet classification unit 1506. 5. Classification [0431]

The packet classification unit 1506 classifies the IP flow and sends it to the IP flow presentation unit 1508 for presentation. [0432]

For existing IP flows, proceed from module 1530 to module 1540 in packet characterization unit 1504. Packets determined by module 1530 to be not too old are placed in a queue of the appropriate class. Then in module 1540, the packet is associated with the existing IP flow. As illustrated in Figure 15A, the packets processed here are considered part of the IP flow known to the system. Therefore, the QoS requirements for this packet are considered to be the same as its IP flow, eliminating the need for QoS processing in modules 1532, 1536, and 1542. In another embodiment, all packets are characterized and classified. Proceed from module 1540 to module 1546 in the IP flow presentation section 1508. [0433]

For new IP flows, go from module 1536 to module 1542 in packet characterization unit 1504. In module 1542, packets are classified into QoS classes by referring to the table of IP flow QoS class table module 1544 in which various types of QoS classes are stored, depending on the QoS requirements of the packet. Similar IP flows (IP flows with similar QoS requirements) can be grouped together in Module 1542. In packet and IP flow classification, QoS class grouping, DiffServ priority marking, and TOS priority marking can be considered. From module 1542, proceed to module 1548 of the IP flow presentation section 1508. 6. IP flow presentation [0434]

The IP flow presentation unit 1508 creates an IP flow packet and presents it to the downlink flow scheduler 604. [0435]

For existing IP flows, proceed from module 1540 to module 1546 in packet classification section 1540. In module 1546, packets are added to the associated existing IP flow queue, that is, the current IP flow queue. From module 1546, proceed to IP flow QoS class queuing processor module 1562 for downlink flow scheduler 604. [0436]

For new IP flows, go from module 1542 to module 1548 in packet classification section 1506. With module 1548, this new IP flow can be initialized for presentation to module 1552. In module 1550, the IP flow QoS class is presented to the frame scheduler 604 and placed in the appropriate class queue. Module 1552 presents the IP flow (especially data packets) and IP flow identifier to the IP flow QoS class queuing processor module 1562 of the downlink flow scheduler 604. 7. Downlink flow scheduler [0437]

Typical logical flow of the downlink flow scheduler 604 in Figure 15B Figure 1560 shows the IP flow QoS class queuing processor module 1562, the MAC downlink subframe scheduler module 1566, the hierarchical class-based priority module 1574, and VPN DEN. Includes data table module 1572, SLA priority data table 1570, CPE IP flow queue depth status processor 1582, and link layer affirmative response processor module 1578. [0438]

The downlink flow scheduler 604 in Figure 15B also includes the following QoS class queues: Class 1, 1564a; Class 2, 1564b; Class 3, 1564c; Class 4, 1564d; Class 5, 1564e; Class 6, 1564f, and so on. MAC downlink subframes, i.e. frame n + 1, 1558b; frame n + 2, 1558c; frame n + 3, 1558d; ... includes frames n + p, 1558k. [0439]

In one embodiment, the downlink flow scheduler 604 is physically located at the location of the radio base station 302, but those skilled in the art can provide the same function at a location remote from the radio base station 302. Is self-evident. [0440]

The downlink flow scheduler 604 can be used to schedule downlink subframes. An entire frame can be divided into an uplink part that transmits an uplink frame (called an "uplink subframe") and a downlink part that transmits a downlink frame (called a "downlink subframe"). .. [0441]

Figure 15B also shows the WAP antenna, radio medium, 290d, RF transceiver subscriber antenna 292d, subscriber CPE station 294d, and subscriber workstation 120d. The WAP antenna 290d and the RF transceiver subscriber antenna 292d are the radio base station 302 (where the downlink flow scheduler 604 is located in one embodiment) and the subscriber CPE station 294d (IP flow is activated by the subscriber workstation 120d), respectively. Provides a wireless connection between the antennas (which can be sent to the application). The WAP antenna 290d acts as the wireless gateway for data network 142, and the RF transceiver subscriber antenna acts as the wireless gateway for subscriber CPE station 294d. This connection is also shown in Figures 2D and 3B. The IP flow QoS class queuing processor module 1562 receives a packet from the IP flow presentation unit 1508. Module 1562 creates class queues 1564a-1564f and places packets in these class queues. Here the number of class queues 1564a-1564f is variable. The method of placing packets on the class queue 1564a-1564f is determined by the input to module 1562. [0442]

Module 1562 can receive input from priority module 1574, VPN DEN data table module 1572, and service level agreement (SLA) priority data table 1570 based on hierarchical classes. Module 1562 features can queue based on these inputs. [0443]

The SLA Priority Data Table 1570 can influence the queuing function with a defined service level agreement for a particular customer. Customers can receive higher quality telecommunications services, or premium services, for example by paying an additional fee. Algorithms running on Module 1562 can increase the queuing priority of messages sent to such customers. [0444]

VPN (virtual private network) DEN (directory enabled networking) data table 1572 can prioritize a defined service quality VPN for companies that pay for VPN functions. VPN is understood by those skilled in the art as a private network provided by a telecommunications service provider with guaranteed bandwidth allocated on the network. The VPN DEN data table 1572 allows module 1562 to provide higher quality of service for customer-purchased VPNs. A queuing priority can be increased for that VPN, as in the SLA Priority Data Table 1570. For example, a class of low-priority IP flows for a platinum-level VPN can be given a higher priority than a class of high-priority IP flows for a brass-level VPN. [0445]

Both the SLA Priority Data Table 1570 and the VPN DEN Data Table 1572 can receive input from the OAM &amp; P (Operation, Management, Maintenance and Conditioning) module 1108. This is a module that is kept offline and includes storing and revising management information about new customers or updating information about existing customers. OAM &amp; P updates customer SLA priorities and VPN information, for example. [0446]

Hierarchical Class-Based Priority Module 1574 operates according to the principle of hierarchical class-based queuing. Hierarchical class-based queuing was created by Sally Floyd and Van Jacobson and is considered the early Internet architecture. [0447]

Queuing based on hierarchical classes uses a tree structure to classify different types of IP flows on edge access device routers. Each branch of the tree represents the IP flow of each class, and each class is dedicated to a limited amount of bandwidth. In this way, the flow of each class so that one IP data flow in one class, or one class in an IP flow, does not run out of all available bandwidth. The minimum bandwidth is guaranteed. The present invention adds a prioritization function that allows class-based priority reservations made using the concept of hierarchical class queues, as already described in FIGS. 13 and 14. [0448]

The MAC downlink subframe scheduler 1566 retrieves packets in class queue 1564a-1564f, reserves frame slots based on priorities 1570, 1572, and 1574, and fills a variable number of subframes 1568a-156k. It is a processor module that can. In one embodiment, each frame is scheduled (filled) with a defined number of packets from each class 1564a-1564f, depending on the priorities 1570, 1572, 1574. In another embodiment, the subframes are scheduled for isochronous reservation by the pre-reservation algorithm method of the invention described with reference to FIGS. 13 and 15. In yet another embodiment, the subframes are scheduled according to a combination of known methods and the pre-booking algorithm method of the present invention. [0449]

The subframe is then transmitted to the WAP antenna 290d, which can be wirelessly transmitted over the radio medium to the RF transceiver subscriber antenna 292d connected to the subscriber CPE station 294d. Subscriber CPE station 294d can then send packets in subframes to subscriber station 120d at CPE subscriber location 306d. Subframes can be scheduled from highest priority to lowest priority [0450]

Hierarchical class-based priority (HCBP) processor module 1574 receives scheduled subframes transmitted from WAP antenna 290d as input. While maintaining the awareness of the packet status (by knowing which packet was sent), the HCBP processor module 1574 knows which class queue 1564a-1564f queues still need to be scheduled. [0451]

Packets can occasionally be lost due to noise, for example. When packet loss occurs, subscriber CPE station 294d sends a retransmission request 1576 to WAP290d, which sends this request to Link Layer Acknowledgment (ARQ) processor 1578. The ARQ processor 1578 notifies the MAC downlink subframe scheduler 1566 of this status, and the MAC downlink subframe scheduler 1566 reschedules and resends the packet requested from the predetermined class queue 1564a-1564f. The link layer acknowledgment (ARQ) processor 1578 also waits for an acknowledgment from subscriber CPE station 294d to determine if the data packet was successfully received. Only after receiving an acknowledgment, the MAC downlink subframe scheduler 1566 deletes the packet from class queue 1564a-1564f. [0452]

Each subscriber CPE station 294d has a limited amount of memory available for data packets received in the IP flow. For example, a device connected to subscriber CPE station 294d (eg, workstation 120d) stops receiving IP data flows when the CPE data packet queue at CPE subscriber station 294d fills up rapidly (eg, subscriber workstation). 120d goes down). In this scenario, subscriber CPE station 294d sends a CPE IP flow queue depth message 1580 indicating that the queue is full. The CPE IP Flow Queue Depth Status Processor 1582 can receive the transmitted CPE IP Flow Queue Depth Message 1580. CPE When the IP flow depth processor 1582 notifies the MAC downlink subframe scheduler 1566 of this status, the MAC downlink subframe scheduler 1566 stops scheduling the downlink subframe to the subscriber CPE station 294d. Processor 1582 can also send a message to the MAC downlink subframe scheduler 1566 to clear certain IP flows from class queue 1564a-1564f. h. Uplink subframe prioritization 1. Overview [0453]

Figures 16A and 16B are typical logical flow diagrams for uplinks. The logical flow is the radio medium up to radio base station 302 for transmission to destination host workstation 136a, and the IP from subscriber workstation 120d connected to subscriber CPE station 294d, which is transmitted to data network 142. For analysis and scheduling of shared radio bandwidth to packet flow. FIG. 16A is a typical logical flow diagram 600 of the uplink IP flow analyzer 632. FIG. 16B is a typical logical flow diagram 1660 of the uplink flow scheduler 634. The functional components of FIGS. 16A and 16B will be described by the method module. Method modules can be thought of as physical units (software, hardware, or combinations thereof) or logical means (used for explanatory purposes only). It is self-evident to those skilled in the art that these modules are used only for purposes of explaining this embodiment and are not limiting. [0454]

A typical logical flow of the uplink IP flow analyzer 632 of FIG. 16A FIG. 1600 includes a packet header identification unit 1602, a packet characterization unit 1604, a packet classification unit 1606, and an IP flow presentation unit 1608. The details of the functions of each part will be described. [0455]

In one embodiment, the uplink IP flow analyzer 632 is physically located at location of radio base station 302, but those skilled in the art may deploy the same function at a location away from radio base station 302. It is self-evident that it is possible. In a preferred embodiment of the invention, the function of the IP flow analyzer 632 is performed at subscriber CPE station 294d, which desires an uplink reserved slot to uplink packet / IP flow to base station 302. A reservation request block (RRB) request detailing the IP flow identifier, the number of packets, and the classification of IP flows is generated by the IP flow analyzer 632 and by the uplink frame scheduler 634, preferably via a competing RRB slot. It can be scheduled to the future uplink subframe slot of station 302. [0456]

Figures 2D, 3A, and 3B help you understand the uplink IP flow analyzer. 2. Introduction [0457]

The IP Flow Analyzer 632 identifies, characterizes, classifies, and presents data packets to the uplink frame scheduler 634. The functions for identifying, characterizing, classifying, and presenting data packets are executed by the packet header identification unit 1602, packet characterization unit 1604, packet classification unit 1606, and IP flow presentation unit 1608 of the uplink IP flow analyzer 632, respectively. To. [0458]

The packet header identifier 1602 indicates whether the packet of the incoming IP flow is part of an IP flow known to the system (ie an existing IP flow) or is the first data packet of a new IP data flow, and the source. Judge the application based on the fields in the packet header. The packet header identification unit 1602 can also buffer, extract, and parse the contents of the header. Packet characterization unit 1604 characterizes a new data packet (or new data flow), determines QoS requirements for that IP data flow based on the source application, and identifies the subscriber workstation that receives the IP data flow. .. The packet classification unit 1606 classifies the new IP data flow into multiple priority classes. The classification unit 1606 can also group packets by similar types of IP flows. IP data flow presentation 1608 initializes a new IP data flow and presents it to the uplink flow scheduler 634. [0459]

Each time the subscriber CPE station 294d attempts to communicate with the radio base station 302 in the uplink direction, it inserts the RRB into the uplink subframe and requests a reservation. The uplink flow scheduler 634 then schedules the reservation request to the future uplink subframe and notifies the CPE station 294d of the reservation. The uplink flow scheduler 634, preferably located at radio base station 302, is a downlink signal that transmits the reserved slot for a particular future frame for the requesting subscriber CPE station 294d and transmits its uplink data. The uplink flow scheduler 634 allocates reservations based on the same parameters that the downlink flow scheduler 604 uses on the downlink. In other words, the uplink flow scheduler 634 determines the reserved slot based on the queue class priority and a set of conventions, and schedules the reservation for uplink transmission from the subscriber CPE station 294d using an advance reservation algorithm or the like. .. The convention is priority processor module 1674 based on hierarchical classes, VPN (virtual private network) Determined based on input to the uplink flow scheduler from DEN (directory enabled) data table 1672, service level agreement (SLA) priority data table 1670. The advance reservation algorithm has been described with reference to FIG. 3. Identification [0460]

The packet header identification unit 1602 identifies the IP flow received from the subscriber CPE station 294d based on the packet header content. [0461]

A stream of packets consisting of packets from various IP flows (each IP flow is associated with one "call") is received by the packet header identification unit 1602. In one embodiment, the IP flow is transmitted from one or more subscriber workstations 120d to the subscriber CPE station 294d for uplinking to the host computer 136a connected to the radio base station 302 by the data network 142. The subscriber CPE station 294d can transmit the IP flow data packet to the packet buffer module 1610 of the packet header identification unit 1602. In one embodiment, the packet header identifier is in the CPE subscriber base station 294d. On module 1610, the received packets are buffered in the storage area for forwarding to the header extraction module 1620. Module 1620 extracts the packet header file and parses it to get the contents of the packet header fields. [0462]

If there are relevant fields, they can include, for example, source, destination, service type (TOS), differentiated services (DiffServ) marking, and so on. [0463]

For IP flows known to the system, there is an entry in the existing IP flow identification table 1626. If the IP flow of an existing IP data call already identifies the previous packet, then the IP flow is in the system. Module 1622 determines if there is a match between the incoming IP flow and the entries in table 1626. If it is determined to match, the IP flow is known to the system and is sent to module 1630 of packet characterization unit 1604. [0464]

If the IP flow is not known to the system, then the IP flow is a new IP data flow and is sent to module 1624 where the packet header fields are analyzed to identify the source application of the IP flow. [0465]

The packet header analysis module 1624 determines from the header table 1628 of the source application the type of source application that is making the data call. The application may be of the type shown in Figure 2D, or one known to those of skill in the art. Examples include FTP (file transfer protocol) downloaded from other client workstations 138f, voice phone calls from callers 124d (connected via a modem), fax calls, multiple callers 124d, 126d (modems). (Connected via), such as a conference call. If the IP flow is a new IP flow, the identity of the new IP flow is added to Table 1626, going from analysis module 1624 to module 1632 in packet characterization unit 1604. 4. characterization [0466]

Packet characterization unit 1604 characterizes a new IP flow and then sends it to packet categorization unit 1606 for classification. [0467]

If the IP flow is an existing IP flow, the process proceeds from module 1622 to module 1630 of the packet header identification unit 1602. If module 1622 determines that the IP data flow is known to the system, module 1630 determines if the packet is too old. This can be done, for example, by determining the packet existence time from the time to live packet header field (IP packet header field) and comparing the field with the threshold lifetime value. If the packet is determined to be too old, it can be discarded. Module 1630 can anticipate dropping application packets. Proceed from module 1630 to module 1640 in packet classification section 1606. [0468]

For the new IP flow, proceed from module 1624 of packet header identification unit 1602 to module 1632. If Module 1624 determines that an application related to an IP flow application is unknown to the system, then the IP Flow QoS Requirements Reference Module 1632 determines the QoS requirements of the application associated with that IP flow. Module 1632 performs this operation with reference to IP Flow QoS Requirements Table 1634. Different applications have different QoS requirements. For example, bandwidth allocation (ie, allocating the right amount of bandwidth) is important for applications that perform FTP file transfer downloads, but jitter (time synchronization of received data), latency (between responses). Elapsed time) is not important. On the contrary, jitter and latency are important for voice calls and conference calls, but bandwidth allocation is not. [0469]

After processing by module 1632, proceed to module 163b. CPE Subscriber Station Identifier (ID) Reference Module 1636 performs a CPE ID reference for the new IP data flow. Each subscriber CPE station 294d may have one or more applications running on one or more subscriber workstations 120d homeed therein. Thus, one or more subscribers can generate or receive an IP flow from the subscriber CPE station 294d, or generate or receive an IP flow at the subscriber CPE station 294d. The subscriber workstation 120d may be a device connected to the subscriber CPE station 294d. Module 1636 references the CPE station identifier of the IP flow in Table 1638 and provides the CPE ID in the reservation request block (RRB). Proceed from module 1636 to module 1648 in packet classification section 1606. 5. Classification [0470]

The packet classification unit 1606 classifies the IP flow and sends it to the IP flow presentation unit 1608 for presentation. [0471]

For existing IP flows, proceed from module 1630 to module 1640 in packet characterization unit 1604. If module 1630 determines that the packet is not too old, module 1640 associates the packet with an existing IP flow. As illustrated in Figure 16A, the packets processed here are considered part of the IP flow known to the system. Therefore, the QoS requirements for this packet are considered to be the same as its IP flow, eliminating the need for QoS processing in modules 1632, 1636, and 1642. [0472]

For new IP flows, go from module 1636 to module 1642 in packet characterization unit 1604. In module 1642, packets are classified into QoS classes according to the QoS requirements of the packet by referring to the table in IP Flow QoS Requirements Table 1644 where the QoS classes are stored. From module 1642, proceed to module 1648 of the IP flow presentation section 1608. 6. IP flow presentation [0473]

The IP flow presentation unit 1608 creates an IP data flow packet and presents it to the flow scheduler 604. In the uplink direction embodiment, a reservation request block (RRB) is generated, uplinked to the radio base station 302 via the connection slot, and scheduled by the IP flow scheduler 634. In another embodiment, the scheduler is located at CPE station 294d so that no reservation request is required. [0474]

For existing IP flows, the packet classification unit 1640 proceeds from module 1640 to module 1646. In module 1646, packets are queued to the IP flow queue, that is, the current existing IP flow queue. In one embodiment, this can include an RRB. Proceed from module 1646 to module 1662 of the uplink flow scheduler 634. In one embodiment, this can include an RRB uplink from CPE294d to radio base station 302. [0475]

For new IP flows, proceed from module 1642 to module 1648 in packet classification section 1606. With IP Flow Initialization Module 1648, this new IP flow is initialized for presentation to Module 1652. Module 1652 presents the IP data flow (especially the reservation request block data packet) to module 1662 of the uplink flow scheduler 634. In module 1650, the QoS class of the IP flow, preferably also including RRB, is presented to scheduler 634. 7. Uplink flow scheduler [0476]

Typical logical flow of the uplink flow scheduler 634 in Figure 16B Figure 1660 shows the IP flow QoS class queuing processor module 1662, the MAC uplink subframe scheduler module 1666, the hierarchical class-based priority module 1674, and VPN DEN. Includes Data Table Module 1672, SLA Priority Data Table 1670, CPE IP Flow Queue Depth Status Processor 1682, Link Layer Positive Response Processor Module 1678. [0477]

The uplink flow scheduler 634 in Figure 16B also has Class 1, 1664a; Class 2, 1664b; Class 3, 1664c; Class 4, 1664d; Class 5, 1664e; Class 6, 1664f QoS class queues and frames n, 1658a. Frame n + 1, 1658b; Frame n + 2, 1658c; Frame n + 3, 1658d; ... Frame n + p, 1658k Includes MAC uplink subframes. [0478]

In one embodiment, the uplink flow scheduler 634 is physically located at the location of the radio base station 302, but those skilled in the art can provide the same function at a location remote from the radio base station 302. Is self-evident. For example, in another embodiment, the uplink flow scheduler 634 can be placed at the location of the CPE station 294d to communicate with the other CPE station 294 and the radio base station 302. [0479]

Uplink flow scheduler 634 can be used to schedule uplink subframes. An entire frame can be divided into an uplink part that transmits an uplink frame (called an "uplink subframe") and a downlink part that transmits a downlink frame (called a "downlink subframe"). .. [0480]

Figure 16B also shows the WAP antenna 290d, the radio medium, the RF transceiver subscriber antenna 292d, the subscriber CPE station 294d, and the subscriber workstation 120d. WAP290d and RF transceiver subscriber antenna 292d are operating radio base station 302 (where uplink flow scheduler 634 is located in one embodiment) and subscriber CPE station 294d (IP flow subscriber workstation 120d), respectively. Provide a wireless connection between (which can be sent upstream to the application). The WAP290d acts as the wireless gateway for data network 142 and the RF transceiver subscriber antenna 292d acts as the wireless gateway for subscriber CPE station 294d to uplink IP flow packet data. [0481]

Figure 16B shows a data interface 320 that provides a connection from the uplink flow scheduler 634 for sending uplink IP flow packets on the data router 140d on data network 142 and to the destination host computer 136a. There is. These connections are also shown in Figures 2D and 3B. [0482]

The previous frame contains an uplink summary request from the subscriber CPE station 294d received by the radio station. At this point, the reservation request block is identified, characterized, classified and presented, preferably on the CPE station 294 side, and is being transmitted from the uplink flow analyzer 632 to the uplink flow scheduler 634 on the CPE 294d. .. In particular, the reservation request block is presented to the IP Flow QoS Class Queuing Processor Module 1662 from Module 1650. Module 1662 notifies the MAC uplink subframe scheduler 1666 of the reservation. [0483]

Conversely, the MAC uplink subframe scheduler 1666 uses slots in subframes to acknowledge a request, called an ARB (Acknowledgement Request Block). Typical slots used to transmit frames, slots, and IP flow identifiers for this reservation have been described with reference to FIG. The scheduler 1666 sends CPE identification data in this reserved slot. Accordingly, the requesting subscriber CPE station 294d can use future slots and frames for the requested data packet IP flow transmission. [0484]

Future slots in future frames are allocated based on input from, for example, priority module 1674, VPN DEN data table module 1672, and service level agreement (SLA) priority data table 1670 based on hierarchical classes. The functionality of these components is similar to the priority module 1574, VPN DEN data table module 1572, and SLA priority data table 1570 based on the hierarchical classes described in the downlink flow scheduler 604. [0485]

When the IP Flow QoS Class Queuing Processor Module 1662 receives a packet of an existing IP or a new IP from the IP Flow Presentation Module 1608, it generates a class queue 1664a-1664f consisting of a variable number of queues and sends the packets to these. Place in the class queue. In a preferred embodiment, the number of classes should be between 3 and 10. These queues hold reservation request packets for scheduling. Packets are placed on class queue 1664a-1664f, depending on the contents of the reservation request block for input to module 1662. [0486]

Module 1662 receives input from priority module 1674, VPN DEN data table module 1672, and service level agreement (SLA) priority data table 1670 based on hierarchical classes. The queuing function of module 1662 is performed based on the inputs from them. The functionality of these components is similar to that of the corresponding components of the downlink flow scheduling method. The SLA Priority Table 1670 and VPN DEN Data Table 1672 can receive input from the OAM &amp; P (Operation, Management, Maintenance and Conditioning) module 1108. The OAM &amp; P module 1108 updates the priority, for example, if the subscriber modifies its service level contract or if the VPN subscriber changes it. [0487]

The MAC uplink subframe scheduler 1666 retrieves requests in class queue 1664a-1664f and schedules reservations for a variable number of slots in frames 1668a-166k. In one embodiment, each frame is scheduled (filled) with a defined number of packets from each class of class 1664a-1664f. Requests can be scheduled with specific priorities in mind, as shown in Figure 13. In another embodiment, it is scheduled by the pre-booking algorithm method of the present invention that schedules the isochronous type of traffic described in FIG. In yet another embodiment, frames are scheduled according to a combination of known methods and the pre-booking algorithm method of the present invention. [0488]

The reserved slot schedule can then be sent to CPE station 294 using FDB slots such as 1236g and 1236h on Figure 12F. The uplink slot is inserted into the uplink subframe by CPE station 294d as scheduled. The frame slot is transmitted from the CPE station 294d to the radio base station 302 and transmitted as a packet to the destination address. For example, a packet from the radio base station 302 can be transmitted to the host computer 136a via the data network 142. [0489]

When the radio base station 302 receives the uplink packet, the radio base station 302 sends an upstream affirmative response data block (UAB) message back to the source subscriber CPE station 294d to acknowledge the reception of the transmitted data packet. .. [0490]

Packets can occasionally be lost due to noise or interference in the radio medium. When packet loss occurs, the subscriber CPE station 294d determines whether it has received the UAB data acknowledgment, and if it determines that it has not received it, WAP290d makes a retransmission request requesting another uplink reserved slot. It sends to radio base station 302 via radio base station 302 and sends a request to link layer acknowledgment (ARQ) processor 1678. The ARQ processor 1678 notifies the MAC uplink subframe scheduler 1666 that it needs to be retransmitted (reservation of a frame slot to retransmit the uplink packet). CPE subscriber station 294d can also send other data messages to the ARQ processor 1678 stating that it has not received an uplink acknowledgment. The ARQ1678 can forward such messages to the uplink subframe scheduler 1666. The uplink subframe scheduler 1666 reschedules the requested uplink reservation from the appropriate class queue 1664a-1664f. Alternatively, in another embodiment, the link layer acknowledgment processor 1678 can send a UAB acknowledgment to the subscriber CPE station 294d to indicate that the data packet was successfully received. This allows the uplink scheduler 1666 to schedule a rebooking for lost packets in addition to scheduling the first (first) reservation. [0491]

Each subscriber CPE station 294d has a limited amount of available to queue packets received from subscriber workstation 120d, waiting for an uplink reserved slot from CPE 294d to radio base station 302. It has a memory space. For example, if the queue at subscriber CPE station 294d is full with backups of packets waiting for upstream reservations, IP data flow may be lost or packets may become out of date. In this scenario, subscriber CPE station 294d sends a CPE IP flow queue depth message 1680 to radio base station 302 indicating that the queue is full. CPE IP Flow Queue Depth Message 1680 is CPE It can be received by the IP flow queue depth status processor 1682. Processor 1682 can notify the MAC uplink subframe scheduler 1666 of this status. The MAC uplink subframe scheduler 1666 temporarily prioritizes the IP flow at subscriber CPE station 294d to eliminate the backlog until the queue depth backlog is reduced to an acceptable level. It is possible to stop sending additional downlink packets to CPE station 294. Processor 1682 can also send a message to the MAC uplink subframe scheduler 1666 to clear the reservation request from subscriber CPE station 294d on class queue 1664a-1664f. 4. TCP auxiliary agent TCP is a reliable transport protocol suitable for traditional networks, where congestion is the primary source of packet loss. However, in networks with built-in wireless links, bit errors can cause significant losses. The wireless environment violates many of the prerequisites that TCP depends on, which causes end-to-end performance degradation. For example, Balakrishman, H .., Seshan, S. and Katz, RH, "Improving Reliable Transport and Handoff Performance in Cellular Wireless Networks", University of California at Barkeley, Berkeley, CA, accessible on the Internet at the following addresses: http: // www. cs. berkeley. edu / -ss / papers / winet / html / winet. html Deales with handoffs and bit error rates in narrowband wireless environments, which is also referenced in this document. Addressing this issue would fix TCP. However, overcoming this challenge is not commercially easy. Solutions that require changes to TCP's standard behavior are impractical. [0492]

The present invention uses an extended MAC layer that interfaces with a TCP auxiliary agent to block TCP layer requests, manipulate the TCP layer at the source or destination of the transmission, TCP operation, and a wireless link in between. Modify with the source or destination of the including TCP / IP transmission. Packets can be queued at a radio base station waiting for an acknowledgment, and the base station overcomes the problem of packet loss due to high bit error rates by performing local retransmissions over the radio link. Can be done. Communication over wireless links is characterized by bandwidth limitations, high latency, sporadic high bit error rates, and temporary disconnects that are addressed by network protocols and applications. There is a need to. [0493]

Reliable transport protocols such as TCP are suitable for traditional preferred networks. TCP handles packet loss due to end-to-end delays and congestion, and is performing very well in such networks. TCP maintains an estimated RTT (Response Time) operating average and average deviation, and provides reliability by retransmitting packets that do not receive an acknowledgment within the average deviation of four times. doing. Wired networks have relatively low bit error rates, so all packet loss is attributed to congestion. [0494]

Although the bit error rate is high in a wireless environment, TCP reacts to packet loss in the same way as in a wired environment. That is, the size of the transmission window is reduced before the packet is retransmitted, a congestion control mechanism or a congestion avoidance mechanism (slow start) is started, and the retransmission timer is reset. These methods unnecessarily reduce link bandwidth utilization, resulting in significant degradation in throughput performance and significant interactive latency. [0495]

The present invention maintains packets waiting in the class queue waiting for an acknowledgment received from the subscriber CPE station. Data slots without an acknowledgment can be retransmitted by having the radio base station perform a local retransmission on the subscriber CPE station. By using a copy of the acknowledgment that identifies the packet loss and performing a local retransmission as soon as the loss is detected, the radio base station can protect the sender from the high bit error rate inherent in the radio link. it can. In particular, even if there is a situation in which the communication quality is very low or a disconnection occurs temporarily during communication, such a situation can be hidden from the sender. [0496]

In data transmission from a CPE subscriber host to a radio base station host, lost packets can be detected by the radio base station and an acknowledgment of loss can be generated for those packets. An acknowledgment of loss can require the CPE subscriber host (sender) to resend the packet. The CPE subscriber host can process the acknowledgment of the loss and retransmit the corresponding loss packet. The present invention has an advantage in that by arranging the TCP sensing type function in the MAC layer, it is not necessary to modify the transmitting side TCP or the receiving side TCP. [0497]

FIG. 5A has a transmitting protocol stack from the transmitting TCP of the subscriber host through the CPE subscriber station on the down side, and a TCP auxiliary agent on the up side via a wireless medium to the radio base station. The flow 500 from the wireless base station's protocol stack to the destination host via the wired connection and the protocol stack is shown. The auxiliary TPCP agent modifies the TCP sliding window algorithm and links it with PRIMMA (Proactive Reservation-based Intelligent Multimedia-aware Media Access) MAC (Media Access Control) on the sending TCP to localize via wireless media according to the present invention. Enables re-transmission. [0498]

In particular, the flow 500 is from the subscriber workstation 120d via the CPE subscriber station 294d at the CPE subscriber location 306d, then via the radio transmission medium via the radio base station 302, and in some cases via a wired link, data network 142. Shows the IP packet flow to the host workstation 136. [0499]

The TCP Auxiliary Agent 510e ensures transport reliability by modifying the behavior of the TCP sliding window algorithm on the sending TCP to optimize the window for the wireless medium. The TCP auxiliary agent 510e has the advantage of being transparent to industry standards protocols and the agent 510e does not require modification of the standard TCP / UDP layer of client subscriber workstation 120d or host workstation 136a. [0500]

Flow 500 is from application layer 512a down, TCP / UDP layer 510a, IP layer 508a, point-to-point (PPP) layer 520a, datalink Ethernet layer 504a, 10-based T Ethernet work interface card (NIC) physical layer 502a. Includes IP flow over a wired connection to the 10-Base T Ethernet Work Interface Card (NIC) Physical Layer 502b of Subscriber CPE294d. [0501]

Subscriber CPE294d sends packets arriving from NIC 502b up its protocol stack to Ethernet layer 504b, PPP layer 520b, 520c and then down to wireless physical layer 502c, including antenna 292d, via PRIMMA MAC 504c. Then, it is sent to the antenna 290d of the radio base station 302 via the radio medium. [0502]

The radio base station 302 sends the packet IP flow received from the antenna 290d on the physical layer 502d to the TCP auxiliary agent 510e via the PRIMMA MAC layer 504d, the PPP layer 520a, and the IP layer 508d in the upward direction. TCP Auxiliary Agent 510e provides IP flow over IP layer 508e, PPP layer 520e, wide area network (WAN) layer 504e, wired physical layer 502e, interface 320, router 140d, data network 142, WAN host workstation 136a over wired connection. Can be sent to the wired layer 502f. [0503]

Host workstation 136a has IP flow from wired layer 502f The protocol stack is sent up to WAN layer 504f, PPP layer 520f, IP layer 508f, TCP / UDP layer 510f, and application layer 512f. [0504]

TCP / UDP layers 510a and 510f provide transport functions such as segmentation, transmission window management, loss packet flow sorting, requesting, and retransmission. [0505]

The TCP layers 510a and 510f usually send a window of packets and then wait for an acknowledgment or retransmission request. The TCP sliding window algorithm is typically used to modify the transmit flow, optimize the transport, and back off when a retransmission request is received and congestion is detected. Unfortunately, in a wireless environment, some packets may not reach the source due to the high bit error rate. Since some of the packets may not arrive at the destination due to the high bit error rate, not due to congestion, the destination IP host prompts the source to resend. Instead of slowing down the transport, the TCP Auxiliary Agent 510e modifies the behavior of the TCP sliding window algorithm to optimize its behavior over the radio. PRIMMA The MAC layer 504d works with the TCP auxiliary agent 510e to request, for example, when the agent blocks a retransmission request from TCP layer 150a on subscriber workstation 120d to host 136a, instead of forwarding the retransmission request to host 136a. Allows the radio base station to send the packet or flow to the subscriber workstation 120d. This is because the packet is not dropped and is still stored in the PRIMMA 504d queue until it receives an acknowledgment from the subscriber CPE. According to the present invention, retransmission can be performed at the PRIMMA MAC data link layer, that is, layer 2, so that retransmission is performed by requesting retransmission to the transmitting TCP via the entire communication path (TCP backs off the sliding window algorithm). It can be done between the base station and the CPE subscribers, rather than (which can cause it to happen). This makes it possible to overcome the problem of high inherent bit error rates while keeping the TCP window optimal by having the radio base station 302 resend until it receives an acknowledgment over the radio link. There is. [0506]

As described above, the TCP transmitter sends a TCP sliding window block of packets and resizes the window when it detects congestion. The TCP transmitter transmits a block of packets in a window and waits for an acknowledgment from the receiver. If the transmission is smooth, that is, if there is no congestion or packet loss, the transmitter TCP will increase the transmission rate. The transmission rate is continuously increased until the sending TCP detects congestion or packet loss. When notified of congestion, TCP stops sending, backs off, and sends packets in smaller blocks (smaller windows). [0507]

The TCP auxiliary agent tricks the sender TCP and its send window algorithm into changing normal TCP behavior. The TCP auxiliary agent, for example, prevents copying of retransmission requests so that the transmitter does not receive loss notifications from the receiving TCP, i.e., congestion notifications. Since the sending TCP does not receive such a notification, it does not change the TCP sliding window and continues transmitting at a high transmission speed. [0508]

When congestion really occurs, that is, when the TCP auxiliary agent recognizes that the packet is really lost, the TCP auxiliary agent can send the retransmission request to the transmitter TCP as it is. This is easy to do because the MAC link layer of the present invention communicates with a high protocol layer and is application sensitive, transport sensitive, and network sensitive. In this case, the MAC layer is a transport layer sensitive type, and the PRIMMA MAC layer 504d communicates with the layer 4 TCP auxiliary agent 510e. Since the MAC requires an acknowledgment of the reception of the radio transmission transmitted to the CPE subscriber station 294d for each packet transmitted from the radio base station 302, the MAC layer 504d is a loss packet in TCP layer communication, for example, radio transmission. Detects whether the retransmission request generated due to loss or real congestion was sent from the CPE station's client computer TCP. [0509]

If the PRIMMA MAC504d does not receive an acknowledgment from the 504c, the PRIMMA MAC504d at radio base station 302 can resend the contents of the lost packet to the subscriber CPE station 294d. If PRIMMA MAC504c at subscriber CPE base station 294d acknowledges reception but still requests retransmission, real congestion can occur. The PRIMMA MAC504d of radio base station 302 can cause the TCP auxiliary agent 510e to detect whether a retransmission request can be sent to the transmitting TCP 510f of the host workstation 136a. [0510]

Thus, the TCP auxiliary agent 510e of the present invention is a TCP sliding window algorithm that optimizes for wireless media without requiring changes to the commercially available TCP layers 510a, 510f used by the receiving and transmitting hosts. You can change the behavior of. In one embodiment, the TCP Auxiliary Agent 510e eliminates the need for modification to the TCP layer on both the sending host and the client. In another embodiment, the host and client TCP layers are unaware of changes in behavior by the TCP auxiliary agent, that is, transparent to the source and destination TCP layers. In another embodiment, the TCP auxiliary agent 510e blocks retransmission requests between the TCP layer of the client computer connected to the subscriber CPE station and the TCP layer of the host workstation connected to the data network. [0511]

FIG. 5B shows a functional flow diagram 522 illustrating a typical function of the TCP auxiliary agent 510e that executes the outgoing TCP spoofing function. As shown in FIGS. 5B and 5A, FIG. 522 assumes that the TCP layer 510f of the sending host 136a has sent a window of packet data to the subscriber workstation 120d and is waiting for an acknowledgment. FIG. 522 assumes that the outgoing TCP message 524 transmitted from the subscriber workstation 120d via the subscriber CPE station 294d is received by the TCP auxiliary agent 510e of the radio base station 302. [0512]

At step 526, the TCP header content of the outgoing TCP message 524 is parsed to reveal the content of the message sent from the subscriber workstation 120d to the sending host 136a over the wireless network. [0513]

At step 528, it is determined whether the contents of the TCP header contain a duplicate acknowledgment message (a copy of the acknowledgment message) from the CPE station. Receiving a duplicate acknowledgment request (a copy of the acknowledgment request) from the CPE subscriber's location can mean message loss on the radio medium, or a true congestion problem. If the TCP packet is determined to be a copy of the acknowledgment message (duplicate acknowledgment message) in step 528, the process proceeds to step 532, otherwise the process proceeds to step 530. [0514]

At step 530, it is determined that true congestion has occurred, that is, it is not a duplicate acknowledgment message due to a wireless link layer retransmission attempt. Therefore, in step 530, the TCP message can pass through the TCP auxiliary 510e without modification and continue to the flow 500 of the TCP layer 510f of FIG. 5A. [0515]

In step 532, since the duplication acknowledgment is recognized in step 528, it is determined whether or not the packet transmission is completed. Step 532 is performed by intercommunication between the TCP auxiliary agent 510e and the PRIMMA MAC layer 504d. This is an example of the interactive behavior between the PRIMMA MAC and the higher layer protocol, shown by line 428 in Figure 4. As shown in Figure 15B, the retransmission request 1576 from the CPE station 294d to the MAC downlink subframe scheduler 1566 warns the scheduler 1566 to retransmit the lost packet of future frame 1568, which is a link layer acknowledgment (ARQ) processor. Since it is received at 1578, the PRIMMA MAC layer 504d can identify whether the transmission of the packet from the radio base station 302 to the CPE station 294 has been completed. If it is determined in step 532 that the packet transmission is complete, the process continues to step 530 described above. However, if it is determined that the packet transmission is not completed, the process proceeds to step 534. [0516]

At step 534, the TCP auxiliary agent 510e discards TCP message 524 because the packet transmission is not complete and the packet can be considered lost on the radio medium. The process proceeds to step 536. [0517]

At step 536, the TCP auxiliary agent 510e can wait for notification from the PRIMMA MAC504d that the link layer has completed the retransmission of the lost packet and has been received by the link layer acknowledgment processor 1578. The process proceeds from step 536 to step 538. [0518]

At step 538, the PRIMMA MAC504d link layer receives an acknowledgment of completion of retransmission and can resume normal TCP message processing. [0519]

In the other step (not shown), the TCP auxiliary agent and the PRIMMA MAC layer set a threshold limit on the number of retransmission attempts, and when the threshold is reached, proceed to step 530 to send a TCP message. It can be passed through without modification. [0520]

FIG. 5C shows a functional flow diagram 540 illustrating a typical function of the TCP auxiliary agent 510e that executes the incoming TCP spoofing function. As shown in FIGS. 5C and 5A, FIG. 540 assumes that TCP layer 510a of the sending subscriber workstation 120d has sent a window of packet data to host 136a and is waiting for an acknowledgment. Figure 544 shows an incoming TCP message 524 sent from host workstation 136a over data network 142 to subscriber workstation 120d on subscriber CPE294d over a radio medium as a TCP auxiliary agent on radio base station 302. It is said that it was received by 510e. [0521]

In step 544, the TCP header content of the outgoing TCP message 524 is parsed to reveal the content of the message sent from host 136a to the sending subscriber workstation 120d over the wireless network. [0522]

At step 546, it is determined whether the contents of the TCP header contain a duplicate acknowledgment message from host 136a. Receiving a duplicate acknowledgment request from the host can mean message loss on the radio medium, or a true congestion problem. If the TCP packet is determined to be a copy of the acknowledgment message (duplicate acknowledgment message) in step 546, the process proceeds to step 550, otherwise the process proceeds to step 548. [0523]

At step 548, it is determined that true congestion has occurred, that is, it is not a duplicate acknowledgment message due to a wireless link layer retransmission attempt. Therefore, in step 548, the TCP message can pass through the TCP auxiliary 510e without modification and continue to the flow 500 of TCP layer 510a of FIG. 5A. [0524]

In step 550, since it was recognized as a duplicate acknowledgment in step 546, it can be determined whether or not the transmission of the packet is completed. Step 550 can be performed by intercommunication between the TCP auxiliary agent 510e and the PRIMMA MAC layer 504d. This is an example of the interactive behavior between the PRIMMA MAC and the higher layer protocol, shown by line 428 in Figure 4. As shown in Figure 16B, a retransmission request 1676 is received by the link layer acknowledgment (ARQ) processor 1678 from CPE station 294d and warns the MAC downlink subframe scheduler 1666 to retransmit the lost packet in future frame 1568. Therefore, the PRIMMA MAC layer 504d can identify whether or not the transmission of the packet from the CPE station 294 to the radio base station 302 is completed. If it is determined in step 550 that the packet transmission is complete, the process continues to step 548 described above. However, if it is determined that the packet transmission has not been completed, the process proceeds to step 552. [0525]

At step 552, the TCP auxiliary agent 510e discards TCP message 524 because the packet transmission is not complete and the packet can be considered lost on the radio medium. The process proceeds to step 554. [0526]

At step 554, the TCP auxiliary agent 510e can wait for notification from the PRIMMA MAC504d that the link layer has completed the retransmission of the lost packet and has been received by the link layer acknowledgment processor 1578. The process proceeds from step 554 to step 556. [0527]

At step 556, the PRIMMA MAC504d link layer can receive an acknowledgment of completion of retransmission and resume normal TCP message processing. [0528]

In the other step (not shown), the TCP auxiliary agent and the PRIMMA MAC layer set a threshold limit on the number of retransmission attempts, and when the threshold is reached, proceed to step 548 to send a TCP message. It can be passed through without modification. 5. Wireless QoS Sensitive PRIMMA MAC (Medium Access Control) Hardware Architecture [0529]

Figure 10 shows the PRIMMA MAC hardware architecture 1000. Architecture 1000 shows a data network 142 connected by a wireless bidirectional connection to WAN interface 320. [0530]

The WAN interface 320 is bidirectionally linked to the bidirectional data frame FIFO1002, which is bidirectionally connected to the SAR (segmentation and requencing) 1004 and the QoS / SLA rules engine and processor 1008. [0531]

The QoS / SLA rules engine and processor 1008 are also bidirectionally connected to the IP flow buffer 1014 and the flash RAM (random access memory) 1010. [0532]

The SAR1004 is bidirectionally connected to the IP flow buffer 1014, flash RAM1010, QoS / SLA rule engine and processor 1008, and PRIMMA MAC scheduler ASIC1012. [0533]

The PRIMMA MAC scheduler ASIC1012 is also bidirectionally connected to RF interface 290, SRAM radiocell buffer 1018, and IP flow buffer 1014. 6. Wireless base station software organization [0534]

FIG. 11 is a diagram of a typical software organization for a packet-centric wireless point-to-multipoint telecommunications system. The software organization in Figure 11 is a wireless transceiver and RF ASIC (aplication specific integrated circuit) module 290, IP flow control 1102, WAN interface management 1104, QoS and SLA management 1106, system OAM &amp; P 1108, customer billing and logging. Includes section 1110, DEN (directory enabled networking) section 1112, and radio base station 320. [0535]

The IP flow control module 1102 includes a transmit queue control module 1102a, a TCP rate control &amp; service class module 1102b, a wireless PRIMMA MAC tier engine 1102c, and an IP flow identification and analysis module 1102d. [0536]

WAN Interface Management Unit 1104 includes WAN In / Out Queue Control Module 1104a, WAN Interface Ports (eg T1, T3, OC3 Ports) 1104b, Firewall and Security Module 1104c, and WAN Traffic Shaping Module 1104d. The IP flow control unit 1102 and the WAN interface management unit 1104 are the "core" parts of the system, and are equipped with packet processing, MAC layer scheduling, TCP proxy agent, and WAN / IF control functions. Many of the functions of the "non-core" parts described above support and control these core parts. The QoS and SLA management unit 1106 includes the QoS performance monitoring and control module 1106a, the service level agreement module 1106b, the policy manager module 1106c, and the encryption management module 1106d. The QoS and SLA management unit 1106 provides the statistical data required by the system when appropriately grouping specific IP flows into QoS classes. Generally, when setting the conditions for installing the system, the service provider will ask the subscriber CPE station SLA, policy-based information (operating time, peak data transmission rate tolerance, etc.), subscriber CPE station 294. Download information related to (remotely). Encryption keys or "strengths" specific to the subscriber CPE station or service provider can also be downloaded. The system OAM &amp; P unit 1108 includes an SNMP proxy client module 1108a for WAP, an SNMP proxy client module 1108b for CPE, and a system operation, management and condition setting module 1108c. The System OAM &amp; P Department 1108 enables remote service staff or equipment to monitor, control, service, modify, and repair the system. You can automatically monitor your system performance level and set traps and traces for your system. Subscriber complaints can be addressed using remote testing and debugging services controlled by System OAM &amp; P Unit 1108. The system OAM &amp; P section 1108's automatic trend analysis feature allows you to monitor system capacity limits and condition additional WAN connections in advance. The customer billing and logging unit 1110 includes accounting log &amp; database management module 1110a, transaction inquiry &amp; processing control 1110b, billing &amp; accounting control module 1110c, and user authentication module 1110d. The customer billing and logging unit 1110 provides the service provider with accounting information, billing information, and transaction information regarding the subscribers of the system. For service providers that charge for usage, they can collect cumulative data on system resource usage. For feature types (eg video conferencing, multicasting, etc.), special billing data can be collected and sent to the service provider. This part controls the availability of the subscriber's system by operating the subscriber authentication function. When a subscriber is authenticated and uses the system, a new subscriber authentication entry is created (remotely) by the service provider. Similarly, the subscriber's access to the system can be denied due to reasons such as delinquency in service charges. Service providers can also remotely query the system for accounting-related transactions. DEN (directory enabled networking) section 1112 is the DEN QoS 1112a module, DEN management &amp; condition setting 1112b module, DEN. Includes IPSEC module 1112c, IP-based VPN control &amp; management module 1112d. [0537]

The DEN unit 1112 provides the service provider with a means of inputting system-related information regarding VPN operation based on the subscriber's DEN. Subscriber VPNs need to be "initialized" and "conditioned" in order for the system to properly allocate system resources to VPN subscribers and to recognize and operate these VPNs. The system uses the data from DEN section 1112 to apply an appropriate priority to the IP flow of a given subscriber. [0538]

The packet-centric radio base station of the present invention supports MICROSOFT, INTEL, and CISCO standard DEN (directory enabled networking), which provides a standard structure for how distributed sites manage IP flows. The present invention prioritizes VPN traffic in an LDAP (lightweight directory access protocol) compliant method that enables remote management, condition setting, and management (LDAP is available from Microsoft, Redpmond, and WA). The invention is also LDAP version 2 compliant. The present invention also complies with RFC 1777, the X.500 standard published by the ITU-T (International Telecommunication Union / Telecommunications Department). [0539]

In one embodiment, DEN provides policy-based network management, IPsec-compatible network security, and IPsec-based VPN. The DEN of the radio base station 302 is planned to be CIM (common information model) 3.0 compliant (when the specifications are completed). Radio base station 302 provides native DEN support and supports directory-based DEN QoS mechanism, precedence./priority/differentiated model (packet marking), including reservation model (eg RSVP queuing per flow). .. Radio base station 302 can plan support for DEN network policies and can support internal QoS and network expansion until DEN is complete. 6. IPsec support [0540]

IPsec has already been introduced with reference to Figure 4. IPsec provides a standard method for packet encryption. In VPN tunnel mode, the entire header can be encoded, or encrypted. In order to make the present invention feasible with packet-centric, QoS-sensitive prioritization, the radio base station needs to be able to analyze the contents of the header field of the packet while identifying the packet / IP flow. Therefore, analysis of unencrypted packets is desired. [0541]

Since the present invention already encrypts the data stream before transmitting the frame over the wireless medium, it is not necessary to use IPsec over the wireless link to perform the encrypted transmission. When a service provider deems it desirable to use IPsec, IPsec can be used to authenticate and securely encapsulate only the header and payload, or just the payload. IPsec is usually integrated into a firewall. If the service provider wishes to implement the invention and IPsec, the invention will be implemented behind a firewall. That is, the firewall can be moved to the radio base station. This allows the IPsec stream to be terminated at a base station that can provide base station access to the packet header field. [0542]

FIG. 17 is a diagram showing an IP flow in the downlink direction including IPsec encryption. Similarly, FIG. 18 is a diagram showing IPsec support in the uplink direction of the present invention. [0543]

FIG. 17 shows the encryption layer traveling down from the source host workstation 136a through a protocol stack that supports IPsec and up through radio base station 302 connected to data network 142 for transmission. Through the wireless link to the subscriber CPE294d, then through the subscriber CPE294d's protocol stack in the up direction, then through the wired connection to the data network 142, and then in the protocol stack in the up direction to join. The downlink flow 1700 to the destination subscriber workstation 120d at the person's location 306d is shown. [0544]

In particular, flow 1700 shows an IP packet flow from host workstation 136a through radio base station 302, then through a wireless transmit link to subscriber CPE294d, and through a wired link with subscriber workstation 120d. ing. [0545]

The host workstation 136a passes the IP flow from the application layer 1712h down through the TCP / UDP layer 1710h, IP layer 1708h, optional PPP layer 1706h, Ethernet layer 1705h, and 10-base T layer 1702h via the data network 142. Proceed to 10-base T-layer 1702g, then through Ethernet 1704g, then up the protocol stack, to optional PPP layer 1706g, IP layer 1708g, 1708h, down the Internet firewall and IPsec security gateway 1706f. In the direction, it sends to WAN layer 1704f, wired layer 1702f, data network 142, and wired physical layer 1702e. [0546]

Wired physical layer 1702e of radio base station 302 brings the IP flow up the protocol stack, through the firewalls of WAN layer 1704e, IPsec security gateway 1706e, IP network layers 1708e and 1708d, and then down the encryption layer 1706d. , PRIMMA MAC layer 1704d, and down from the wireless link to subscriber CPE294d. [0547]

Subscriber 294d brings packet IP flow up from antenna 292d on physical radio layer 1702c and down via MAC layer 1704c, encryption layer 1706c, IP layer 1708b, 1708c, optional layer 1706b, Ethernet layer 1704b, 10 Send with base T connection 1702b, 10 base T connection. [0548]

The subscriber workstation 120d sends the IP flow from the 10-base T-layer 1702a up its protocol stack to Ethernet layer 1704a, optional PPP layer 1706a, IP layer 1708a, TCP / UDP layer 1710a, and application layer 1712a. [0549]

FIG. 18 shows the transmission protocol stack from the source TCP of the subscriber workstation 120d at the CPE location 306d to the Ethernet connected to the CPE subscriber station 294d and the radio base station 302 in the down direction, and IPsec. The IP flow in the uplink direction is shown, which advances the protocol stack of the radio base station 302 that supports the above, then proceeds to the wired connection to the data network 142, and reaches the protocol stack to the destination host. [0550]

In particular, Flow 1800 shows an IP packet from the subscriber workstation 120d through the subscriber CPE294d, then over the radio transmission medium to the radio base station 302, and in some cases over the wired link to the host workstation 136a. ing. [0551]

Flow 1800 is from application layer 1812a down the protocol stack TCP / UDP layer 1810a, IP layer 1808a, optional point-to-point (PPP) layer 1806a, datalink Ethernet layer 1804a, 10-based T Ethernet network interface card ( NIC) Physical layer 1802a, including IP flow going through the subscriber CPE294d's 10-base T Ethernet NIC physical layer 1802b over a wired connection. [0552]

The subscriber CPEd advances the packet arriving from NIC1802b up its protocol stack to IP layers 1808b and 1808c via the optional PPP layer 1806b and down the Internet firewall and IPsec security gateway 1806b. Proceed to the radio physical layer 1802c including the PRIMMA MAC1804c and the antenna 292d, and to the antenna 290d of the radio base station 302 of the radio physical layer 1802d via radio media such as RF communication, cable RF, satellite link. [0553]

The radio base station 302 advances the IP flow upward from the antenna 290d of the physical radio layer 1802d to the MAC layer 1804d, the IPsec layer 1806d, and 1806d, and can encapsulate and encrypt the packet. The IP flow from the IPsec layer 1806e goes down to the wired physical layer 1802e via the WAN layer 1804e and the data network 142. [0554]

Wired physical layer 1802f sends the IP flow up the protocol stack through the firewalls of WAN layer 1804f, IPsec security gateway 1806f, IP network layers 1808f and 1808g, and then down the optional PPP layer 1806h, Ethernet layer 1804h, 10 base T layer 1802g in the down direction, interface 320, It is sent to the data network 142 via router 140d and to the 10-base T physical layer 1802h of host workstation 136a via a wired connection. [0555]

Host workstation 136a has IP flow from 10 base T layer 1802h, protocol stack up, Ethernet layer 1805h, Proceed with optional PPP layer 1806h, IP layer 1808h, TCP / UDP layer 1810h, application layer 1812h. IV. Conclusion [0556]

Although various embodiments of the present invention have been described, these examples do not limit the present invention. The scope of the present invention is not limited to the above-described embodiment, but is defined only according to the following claims.

[Simple explanation of drawings]

[Fig. 1A]

FIG. 6 is a block diagram showing an overview of a standard telecommunications network that provides local telecommunications services within one or more local access and transmission areas. [Fig. 1B]

It is a figure which shows an example of the network including the workstation connected to the data network. [Fig. 1C]

It is a diagram of a conventional video network such as a cable television (CATV) network. [Fig. 2A]

It is a block diagram which outlines a telecommunications network which provides a standard telecommunications service between a local telecommunications service provider and a long-distance telecommunications service provider between subscribers located in different local access and transmission areas. .. [Fig. 2B]

It is a figure which shows the detail of a signal network. [Fig. 2C]

It is a figure which shows the typical network which carries voice, data, and video traffic through a data network. [Fig. 2D]

It is a figure which shows the network including the point-to-multipoint wireless network connected to the data network through a router. [Fig. 3A]

It is a typical perspective view of a point-to-multipoint network. [Fig. 3B]

It is a block diagram which shows a wireless point-to-multipoint network. [Fig. 4]

It is a figure which shows the wireless Internet protocol network access architecture of this invention. [Fig. 5A]

It is a figure which shows the Internet protocol flow from a subscriber host to a wireless base station, and from a wireless connection to a destination host. [Fig. 5B]

It is a function flow diagram which includes the explanation of the typical function of the transmission control protocol auxiliary agent which executes the transmission transmission control protocol spoofing function. [Fig. 5C]

It is a function flow diagram which includes the explanation of the typical function of the transmission control protocol auxiliary agent which executes an incoming call transmission control protocol spoofing function. [Fig. 6]

It is a block diagram which shows the scheduling when a plurality of Internet protocol flows are mixed. [Fig. 7]

It is a figure which shows the packet header field information which can be used to distinguish the quality of the service requirement of the Internet Protocol flow and the Internet Protocol flow. [Fig. 8A]

It is a block diagram summarizing typical downlink analysis, prioritization and scheduling functions. [Fig. 8B]

It is a block diagram summarizing typical uplink analysis prioritization and scheduling functions. [Fig. 9]

It is a figure which shows the method in which the downlink flow scheduler performs scheduling by prioritizing and allocating resources to a frame slot in consideration of a service level contract. [Fig. 10]

It is a figure which shows one example of the medium access control hardware architecture of this invention. [Fig. 11]

It is a figure which shows the typical software organization of the packet-centered wireless point-to-multipoint telecommunications system. [Fig. 12A]

It is a figure which shows the typical time division multiplex access medium access control airframe. [Fig. 12B]

It is a figure which shows the typical time division multiple access / time division double air frame. [Fig. 12C]

It is a figure which shows the typical downstream transmission subframe. [Fig. 12D]

It is a figure which shows the typical upstream acknowledgment block field of a downstream transmission subframe. [Fig. 12E]

It is a figure which shows the typical acknowledgment request block field of a downstream transmission subframe. [Fig. 12F]

It is a figure which shows the typical frame descriptor block field of a downstream transmission subframe. [Fig. 12G]

It is a figure which shows the typical downstream medium access control payload data unit of a downstream transmission subframe. [Fig. 12H]

It is a figure which shows the typical command and control block of a downstream transmission subframe. [Fig. 12I]

It is a figure which shows the typical upstream transmission subframe. [Fig. 12J]

It is a figure which shows the typical downstream acknowledgment block of the upstream transmission subframe. [Fig. 12K]

It is a figure which shows the typical reservation request block of the upstream transmission subframe 1204. [Fig. 12L]

It is a figure which shows the typical medium access control payload data unit of an upstream transmission subframe. [Fig. 12M]

It is a figure which shows the typical operation data block of the upstream transmission subframe. [Fig. 12N]

It is a figure which shows the typical operation data block of the upstream transmission subframe. [Fig. 12O]

It is a figure which shows the typical operation data block of the upstream transmission subframe. [Fig. 13]

It is a figure which shows the method of functioning the typical flow scheduler of this invention. [Fig. 14]

It is a typical two-dimensional block diagram of the advance reservation algorithm. [Fig. 15A]

It is a typical logic flow diagram of a downlink flow analyzer. [Fig. 15B]

It is a typical logical flow diagram of the downlink flow scheduler. [Fig. 16A]

It is a typical logic flow diagram of an uplink flow analyzer. [Fig. 16B]

It is a typical logical flow diagram of the uplink flow scheduler. [Fig. 17]

It is a figure which shows the Internet Protocol flow in the downlink direction including the Internet Protocol security encryption. [Fig. 18]

It is a figure which shows the Internet Protocol security support in the uplink direction.