Terminal and method for transporting data packets via radio frames

ES 2 219 019 T3

DESCRIPTION

Terminal and procedure for transporting data packets through radio frames.

Field of the invention

The invention relates to a terminal for a wireless network for a metropolitan area. More particularly, the invention relates to a terminal for transporting Ethernet data packets through radio frames in a metropolitan area wireless network.

Background of the invention

Computers used in modem office environments are typically coupled to a local area network (LAN). The LAN network allows computer users to share common resources, such as a common printer included on the network, and allows users to share information files, such as including one or more file servers on the network. In addition, users can usually exchange information with each other through electronic messaging. One commonly used type of LAN is Ethernet. Currently, a variety of products that support Ethernet are commercially available from a variety of sources. Other types of LAN networks are also used, such as token ring, fiber optic data distribution network (FDDI) or asynchronous transfer mode (MTA).

LAN networks are often connected to a wide area network (WAN) through a telephone modem. In this way, the information is communicated through the WAN network by means of a communication link provided by a telephone service provider. These telephone links, however, are normally designed to have a bandwidth that is sufficient for voice communication. As such, the speed at which information can be communicated over these telephone links is limited. However, as computers and computer applications become more sophisticated, they tend to generate and process larger and larger amounts of data to communicate. For example, computer graphics communication typically requires a large amount of bandwidth relative to voice communication. In this way, the telephone link can become a data communication bottleneck.

Business organizations and their affiliates typically span multiple locations in a metropolitan or geographic area. For example, a business organization may have a central office, one or more branch offices, and various other facilities. For these types of business organizations, the LANs located at the various sites will generally need to communicate information with each other. Wireless communication links for connecting local area networks are known. For example, US Patent No. 4,876,742, entitled "Apparatus and Method for Providing a Wireless Link Between Two Area Network Systems", and US Patent No. 5,436,902, entitled "Ethernet Extender" each describe a link. Wireless communication to connect LAN networks.

Availability is a measure of the average number of errors that occur in digitally transmitted data. 99.99 percent availability is normally required for radio communications. For 99.99 percent availability, the average error rate for digitally transmitted data should be kept below 1 x 10^-6 errors per bit 99.99 percent of the time. However, the integrity of a wireless communication link is highly dependent on transient environmental conditions, such as rainfall. Environmental precipitation causes severe attenuation of the transmitted signal. For example, to maintain 99.99 availability in the presence of environmental precipitation, the signal must be transmitted at a level that is 24 dB / km higher than in the absence of it. Therefore, to ensure an acceptable data error rate under whatever conditions can be expected, data is normally transferred over a wireless communication link at relatively high power and relatively low transmission speed. However, the amount of data that is required to be transmitted over the wireless link can vary greatly over time and can vary regardless of environmental conditions. Additionally, wireless links, especially those operating at high power levels, can cause interference with other wireless links operating in the same geographic area. In this way, the wireless link can become a data communication bottleneck.

Therefore, a technique is needed to communicate data efficiently and cost-effectively over a wireless link between Ethernet local area networks.

Known wireless transmission systems for LAN networks suffer from a disadvantage in that they require conversion of the LAN protocol to an intermediate protocol prior to wireless transmission. Known systems of this type carry out the conversion to a telephony protocol or an asynchronous transfer mode protocol (MTA).

Therefore, what is required is a technique for communicating data over a wireless link between local area networks that is not affected by these drawbacks.

The patent abstract of Japanese Publication No. 07336367 describes a radio communication system aimed at obtaining high transmission efficiency and maintaining a high ability to correct errors proportionally.

ES 2 219 019 T3 by operating a function that performs error correction coding for received data.

US-A-5436902 describes a telecommunications bridge system that uses synchronized arbitration in a part of a frame of a radio frequency system and is transparent to Ethernet layer protocols and Ethernet-type packets to increase utilization. of bandwidth.

Goldberg L: '100BASE-T4 TRANSCEIVER SIMPLIFIES ADAPTER, REPEATER, AND SWITCH DESIGNS' ELECTRONIC DESIGN, vol. 43, no. 6, March 20, 1995 (March 20, 1995), pp. 155-156, 158, 160 XP000509380 ISSN: 0013-4872, describes a 100Base-T4 transceiver that simplifies adapter, repeater and switch designs.

Summary of the invention

According to the present invention, in a first aspect a terminal is provided for transporting data packets through radio frames, the terminal comprising:

a data packet receiver for receiving data packets for communication over a wireless link in which each data packet can be of different length;

a data packet formatting apparatus coupled to the data packet receiver, providing for data packet formatting to dynamically format data packets according to radio frames in which each radio frame has the same length, and a length of a specific data packet is less than, equal to or greater than the length of the radio frame, and wherein each radio frame includes a data field having a predetermined length to receive the data packets, the data packets are received by the receiver of data packets separated by a separation between packets, and a code representative of the spacing between packets is stored in the data field between the data packets; and a wireless transceiver coupled to the apparatus for formatting the packets, the wireless transceiver being for communicating the radio frames over the wireless link.

According to the present invention, in a second aspect there is provided a method for transporting data packets through radio frames in which the method comprises the following steps:

(a) receive data packets in which not all data packets have the same length, and receive a separation between packets between every two contiguous packets;

(b) format the data packets according to radio frames in which the radio frames are each the same length, and a length of a specific data packet is less than, equal to or greater than the length of the frame of radio, each radio frame includes a data field that has a predetermined length, and wherein the shaping comprises placing the data packets one after another in the data field and inserting a code representative of a packet spacing between two contiguous packets; and (c) communicating the radio frames over the wireless link.

Preferred embodiments of the aforementioned terminal and method are defined in the appended dependent claims 2 to 12 and 14 to 32, respectively.

The present invention provides an improvement in that the conversion of the LAN protocol to an intermediate protocol is not required before wireless transmission. Instead, the present invention transfers data packets over a wireless link in a highly efficient manner. Thus, according to the present invention, no conversion is required to convert the LAN protocol into a telephony communication protocol, such as PDH (for example, ED1, ED3, E1 and E3) or SDH (for example, OC-1 , OC-3), or in an asynchronous transfer mode protocol (MTA) prior to communication over a wireless link.

Brief description of the drawings

Figure 1 illustrates a schematic block diagram of a pair of wireless terminals communicating with each other over a wireless communication link in accordance with the present invention.

Figures 2A-F illustrate representative metropolitan area network (MAN) topologies in accordance with the present invention.

Figure 3 illustrates a schematic block diagram of a single wireless terminal 100 in accordance with the present invention.

Figure 4 illustrates a schematic block diagram of the CAM unit for digital signal processing and the radio frame configurator included in the CODEC illustrated in figure 2.

ES 2 219 019 T3

Figure 5 illustrates a frame structure for 100BASE-T Ethernet data packets according to the present invention.

Figure 6 illustrates a radio frame according to the present invention.

Figure 7 illustrates a radio superframe according to the present invention.

Figure 8 illustrates a schematic block diagram of a symbol cipher in accordance with the present invention.

Figure 9 illustrates a schematic block diagram of a differential encoder and characteristic equations according to the present invention.

Figure 10 illustrates a schematic block diagram of a differential decoder and characteristic equations according to the present invention.

Figure 11 illustrates a mapping constellation for a constellation mapper in accordance with the present invention.

Figure 12 illustrates a schematic block diagram of a radio frame Ethernet sync portion of the baud rate control logic in accordance with the present invention.

Figure 13 illustrates a schematic block diagram of a radio-to-Ethernet frame synchronization portion of the baud rate control logic in accordance with the present invention.

Figure 14 illustrates a schematic block diagram of a microwave module and microwave antenna according to the present invention.

Figure 15 illustrates a perspective view of the microwave antenna and a housing for the outdoor unit according to the present invention.

Figure 16 illustrates a schematic block diagram of an alternative embodiment of the digital signal processing CAM unit and radio frame configurator according to the present invention.

Figure 17 illustrates a frame structure for reshaped 100BASE-T Ethernet data packets, formed by the CAM and radio frame shaper of Figure 14.

Figure 18 illustrates a schematic block diagram of an adaptive countermeasures block in accordance with the present invention.

Figure 19 illustrates a graph of received signals level versus time as a result of rain fading.

Figure 20 illustrates a flow chart for implementing countermeasures in accordance with the present invention.

Figure 21 illustrates a point-to-multipoint metropolitan area network divided into sectors having internal and external radii in accordance with the present invention.

Figure 22 illustrates a wireless link between two terminals in which an unauthorized terminal is attempting to eavesdrop on the communication between two terminals.

Fig. 23 illustrates an embodiment according to the present invention having multiple digital processing CAM units multiplexed to a single radio frame shaper.

Detailed description of the preferred embodiment

Figure 1 illustrates a schematic block diagram of a pair of wireless terminals 100, 100 'communicating with each other via a two-way wireless communication link 102 in accordance with the present invention. Although a single wireless communication link 102 is illustrated, it will be apparent that a network of wireless communication links can interconnect a plurality of wireless terminals, thus forming a metropolitan area wireless network (MAN) in accordance with the present invention. Figures 2A-F illustrate representative MAN topologies interconnecting wireless AE nodes with wireless links in accordance with the present invention. Each of the nodes AE may include a wireless terminal identical to the terminal 100 or 100 'illustrated in Figure 1 to terminate each wireless link. It will be apparent that other MAN topologies can be implemented and that one or more of the AE nodes can be coupled to one or more different types of networks.

Due to the availability of portions of the radio spectrum in the 38 GHz frequency band, the wireless link 102 illustrated in FIG. 1 preferably operates within this frequency band, although another frequency band may be selected. Different channels within the selected band are assigned to nearby wireless links to reduce interference between them. The channels are preferably staggered in

ES 2 219 019 T3 25-50 MHz intervals. Since the 38 GHz radio frequency band is susceptible to rain fading, the manner and path of transmissions over the wireless link 102 are adaptively modified to maintain a consistent predefined transmission quality on the network in accordance with the teachings of the master application, serial number 08 / 950,028, filed on October 14, 1997, the contents of which are referred to herein.

Referring to Figure 1, wireless link 102 preferably includes a primary radio channel 102A carrying 100 megabits per second (Mbps) full-duplex data traffic, including payload data, and an auxiliary radio channel 102B. carrying full duplex control data for network management and control over the manner of transmission over link 102 (link control). For example, changes in transmission mode initiated through link control may include changes in transmission power, data bit rate, amplitude modulation scheme, spectrum spread, and path. of transmission.

Terminal 100 includes a broadcast device, also referred to herein as an external unit (UE) 104, that terminates one end of wireless link 102. In the preferred embodiment, the UE 104 includes a two-way radio antenna and is mounted outside on a roof mast of a building. Also included in terminal 100 is an extender device, also referred to herein as upper deck unit 106 (UPS), which is coupled to the UE via two-way communication cables 108, 110, and 112. by power lines 114. The UPS 106 is preferably located inside the building that has the UE 104 positioned on its roof, and as close to the UE 104 as is practical. In the preferred embodiment, the UPS 106 is located indoors, ideally in a wiring closet, on the top floor of the building. It will be apparent that the term "top floor unit" as used herein refers to the extender unit 106 and its equivalents, regardless of its location in relation to a building. For example, the "top floor unit" is preferable but not necessarily located on the top floor of a building.

Cable 108 carries full duplex data traffic between UE 104 and UPS 106, which is received from or transmitted onto primary radio channel 102A. The data traffic communicated over cable 108 includes payload data for communication over link 102 and may also include network management and control data. Preferably, the data communicated over cable 108 conforms to a Fast Ethernet standard, 802.3u, adapted by the Institute of Electrical and Electronics Engineers (IEEE), such as the 100BASE-TX standard or the 100BASE-T4 standard, which operate at a data rate of 100 Mbps. Cable 110 carries half duplex data for network control and management between UE 104 and UPS 106. Preferably, the data communicated over cable 110 conforms to an Ethernet standard, such as the 10BASE-T standard, operating at 10 Mbps. Cable 112 carries serial data for tuning and maintenance purposes between UE 104 and UPS 106. Preferably, the data communicated via cable 112 conforms to the conventional RS423 serial port communication protocol. Cable 114 provides power to UE 104.

Thus, in the preferred embodiment of the present invention, data is communicated between the UPS 106 and the UE 104 through each of the cables 108, 110 and 112 according to baseband communication frequencies. This is in contrast to systems that communicate data between an indoor unit and an outdoor unit by modulating this data at intermediate frequencies (IF). The baseband communication aspect of the present invention has an advantage over such an IF modulation scheme because the present invention simplifies the implementation of the UPS 106. Furthermore, the cables 108, 110 and 112 can be of less expensive construction than the one that would be required for communication in FI.

A router or switch 116 is coupled to the UPS 106 and, therefore, to the terminal 100, through cables 118 and 120. The cable 118 preferably communicates data according to the Fast Ethernet 100BASE-TX or T4 standard, while the cable 120 preferably communicates data according to the 10BASE-T Ethernet standard. Alternatively, the cable 118 may be a fiber optic cable, in which case it preferably communicates data according to the Fast Ethernet 100BASE-FX standard.

A cable 122 is coupled to a serial port of the UPS 106. Preferably, the data communicated over the cable 122 conforms to the RS232 serial port communication protocol. A diagnostic station 124 may be coupled to cable 122 to perform diagnostics, adjustments, and maintenance of terminal 100. Since certain aspects of the UPS 106 and UE 104 can only be accessed from the diagnostic station 124, security in relation to these aspects is enhanced by the requirement that the diagnostic station 124 be directly connected to the UPS. 106 through cable 122. AC power is supplied to UPS 106 through a power cable 126.

A wired local area network (LAN) 128, such as an Ethernet LAN located inside the building having terminal 100, may be coupled to router or switch 116. In addition, a wide area network (WAN) 130, such as a telephone service network that provides access to the Web (World Wide Web), it may be coupled to the 128 LAN. In this manner, the wireless link 102 can be accessed from one or more personal computers (PCs), data terminals, workstations or other conventional digital devices included in the network 128 LAN or the network 130 WAN. A system 132 for network management (SGR) is coupled to any one or more of router or switch 116, LAN 128, or WAN 130. SGR 132

ES 2 219 019 T3 accesses wireless link 102 and terminals 100, 100 'to perform network management and link control functions (for example, collect data related to the operation of the MAN network or change the transmission mode of data through a specific link or links). If the SGR 132 is coupled to the 128 LAN, this access is carried out through the 128 LAN. However, if the SGR 132 is coupled to the WAN 130, this access is remote through direct dialing through a telephone service provider or through Web access. When accessing the network management and link control functions via the web, a web browser is provided on the SGR 132, while a web server 236 (FIG. 3) is provided on the terminal 100. In the preferred embodiment, the ED 124 and SGR 132 are each a personal computer, but may be another type of conventional digital device.

Terminal 100 'terminates the opposite end of link 102, distant from terminal 100. In the preferred embodiment, link 102 can be up to 4 kilometers or more in dry climates (eg, Wyoming), while retaining 99 99% link availability, and can be up to 1.2 kilometers or more in more humid climates (eg Florida), while conserving 99.99% link availability. Elements illustrated in Figure 1 having a one-to-one functional correspondence are given the same reference number, but are distinguished whether the reference number is prime or not. Note, however, that because any SGR 132, 132 'can access the wireless communication link 102 and the two terminals 100, 100', it is not necessary to locate an SGR 132 or 132 'at each end of link 102.

Figure 3 illustrates a schematic block diagram of a single wireless terminal 100, including a UPS 106 and a UE 104, in accordance with the present invention. UPS 106 includes a 100BASE-T regenerator 200 that is coupled to cable 118 (Figure 1) and cable 108 (Figure 1). Also, assuming that the cable 118 is a fiber optic cable, the UPS 106 includes a converter 202 for the conversion between the fiber optic cable and a Category 5 twisted pair cable. Converter 202 is coupled to fiber optic cable 118 and regenerator 200. UPS 106 also includes a 10BASE-T repeater 204 coupled to cable 120 (Figure 1) and cable 110 (Figure 1). A converter 206 included in the UPS 106 converts between signals according to the RS232 standard and signals according to the RS423 standard. Converter 206 is coupled to cable 122 (Figure 1) and cable 112 (Figure 1).

The USO 106 also includes an alternating current to direct current (AC / DC) power transformer 208 coupled to cable 126 (Figure 1) and cable 114 (Figure 2). Power transformer 128 provides power to UPS 106 and UE 104. A status indicator 210 included in UPS 106 displays the status of UPS 106 via light emitting diodes for diagnostic, adjustment, and maintenance purposes.

The UPS 106 provides three interfaces to customer equipment, including the router or switch 116 (Figure 1) and the DI 124 (Figure 1). These include a 100 Mbps full duplex interface through configurator 200, a 10 Mbps half duplex interface through repeater 204, and an RS232 serial port through converter 206. Although payload data traffic is typically directed through the 100 Mbps interface, while traffic for network management and link control is typically directed through the 10 Mbps interface, a user of the Terminal 100 can combine signals for network management and link control with payload data traffic on the 100 Mbps interface depending on the specific capacity of the router or switch 116 (FIG. 1).

The UPS 106 provides an interface from multiple internal cables 118, 120, 122, 126 to multiple external cables 108, 110, 112, and 114. UPS 106 also regenerates / repeats Ethernet signals in the form of Ethernet data packets between cables 108, 118 and between cables 110, 120. In this way, UPS 104 serves to extend the maximum possible distance between customer equipment. , such as router or switch 116 (Figure 1), and UE 104. In the preferred embodiment, a distance between customer equipment and UPS 106 can be up to 100 meters, while a distance between UPS 106 and UE 104 can also be up to 100 meters. Accordingly, in the preferred embodiment, a distance between the customer's equipment and the UE 104 can be up to 200 meters. Since data is communicated between UPS 106 and UE 104 at baseband frequencies, no apparatus is required in UPS 106 that performs IF modulation.

UE 104 includes a 212 100BASE-T transceiver coupled to cable 108, a 214 10BASE-T transceiver coupled to cable 110, a 216 RS423 controller coupled to cable 112, and a DC to DC power transformer 218 coupled to cable 114 The 100BASE-T transceiver 212, 10BASE-T transceiver 214, and RS423 controller 216 are each coupled to an encoder / decoder 220 (CODEC) included in UE 104. Power transformer 218 provides power to UE 104.

The CODEC 220 includes a media access control (CAM) unit 222 having a transmitting part 224 and a receiving part 226, a radio frame configurator 228, and a microprocessor 230 to control the operation of the UE 104. The transmit portion 224 and the receive portion 226 of the CAM 222 are coupled to the 100BASE-T transceiver 212 to communicate Ethernet data packets with the 100BASE-T transceiver 212. Radio frame configurator 228 is coupled to CAM 222 to translate data from Ethernet data packets received by CAM 222 into radio frames 350 (FIG. 6) suitable for radio frequency modulation and transmission. Radio frame configurator 228 also translates received radio frames 350 (FIG. 6) into packets that it provides to CAM 222.

ES 2 219 019 T3

The microprocessor 230 is programmed by software to implement a TCP / IP stack 232, a link management (GE) task 234, a hypertext transfer protocol (HTTP) server 236, and an agent 238 for the simple protocol management protocol. networks (SNMP). Microprocessor 230 manages each wireless link in a network of such wireless links (eg, a MAN network), including a local link 102 (FIG. 1) that is directly coupled to terminal 100. The microprocessor 230 can be accessed through any of the SGRs 132 (figure 1) and through the ED 124 (figure 1). In this way, the wireless link network can be managed locally, such as through an SGR 132 or a ED 124 that are connected to the UPS 106. For this purpose, the microprocessor 230 is assigned a CAM address (control of media access) from Ethernet. Alternatively, the wireless link network can be managed remotely, such as through an SGR 132 that is coupled to the WAN network (figure 1) and that accesses the microprocessor 230 through an Internet access that uses the TCP / IP protocol (Internet protocol). The 232 TCP / IP stack provides for this TCP / IP interface over the Web. For this purpose, the microprocessor 230 is assigned an Internet Protocol (IP) address.

Task 234 GE provides a function to change the way data is transmitted over a wireless link, initiated by one of the SGRs 132, 132 '. For example, the data rate for link 102 can be modified via task 132 GE included in UE 104. This may include sending a command for link control over link 102 to UE 104 '(Figure 1) so that the two terminals 100, 100' communicate data at the same transmission rate. Such commands are received from the microprocessor 230 and provided to it by a module 240 for managing the header link (GE / EN) included in the radio frame configurator 228. In this way, the radio frame configurator 228 suitably combines the traffic for network management and link control, provided by task 234 GE, with payload data, received from CAM 222, in frames 350 of radio (figure 6) for communication on link 102. In addition, radio frame configurator 228 extracts traffic for network management and link control from radio frames 350 (FIG. 6) received from link 102 and supplies it to task 234 GE of microprocessor 230 through the 240 GE / EN module. Although two types of data traffic (payload and link control) are communicated over radio frames 350 (Figure 6), payload data is considered to be communicated over primary channel 102A, while considers that the traffic for the control of the link is communicated through the auxiliary channel 102B. Correspondingly, these two channels 102A and 102B are time division multiplexed.

A graphical user interface, through which the microprocessor 230 can be accessed from an SGR 132, 132 '(figure 1) or a DI 124, 124' (figure 1) for network management and link control, is achieved preferably by HTTP web server software module 236, which is implemented by microprocessor 230 located in UE 104 and assigned a unique IP address. The server software 236 operates in conjunction with the TCP / IP stack 232. In accordance with this aspect of the invention, the server software 236 is used to provide a graphical user interface through which network management functions are initiated. These functions include retrieving data representative of network conditions on the MAN network and changing the way data is transmitted over a wireless link on the MAN network.

In this way, the functions to manage the MAN network and its wireless links can be accessed and these can be initiated from the network management (SGR) stations 132, 132 'located in various parts of the MAN network, using browser software. Resident web on SGR 132, 132. This graphical user interface provides an easy-to-use environment that can run on, and be accessed by, a variety of different SGRs sourced from a variety of different manufacturers. For example, an SGR 132, 132 'can be a workstation manufactured by Sun Microsystems, a personal computer manufactured by any one of a variety of manufacturers, or even a set-top-box used in conjunction with a television set. Compatibility with the Web server is achieved through commercially available Web browser software, resident in SGR 132, 132 '. This aspect of the present invention deals with compatibility issues between SGR 132, 132 'and terminal 100, 100'.

The SNMP agent 238 located in the UE 104 maintains a management information database (BIG statistics), which is a collection of managed objects that correspond to resources of the MAN network and of the terminal 100. The SNMP agent 238 can access the BEG to control certain aspects of the MAN network and terminal 100, and may query the BIG for information related to managed objects. SNMP is accessible through HTTP server 236.

UE 104 also includes a transmit modulator 242 (mod TX), a receive demodulator 244 (demod RX), and a microwave module 246 (MMO). Transmission module 242 translates baseband broadcast digital data received from radio frame configurator 228 into suitable analog waveforms for upconversion to microwave frequencies and final transmission over wireless link 102. The analog waveforms formed by the transmit modulator 242 preferably modulate a 490 MHz IF carrier. However, it will be apparent that a frequency other than 490 MHz may be selected for this purpose.

The receive demodulator 244 performs functions that are essentially the opposite of those performed by the transmit modulator 242. In the preferred embodiment, the receive demodulator 244 receives a 150 MHz IF signal from the microwave module 246. However, it will be apparent that a frequency other than 150 MHz can be selected for this purpose. The receive demodulator 244 controls the level of this signal through automatic gain control (AGC) and then downconverts the signal to the baseband of

ES 2 219 019 T3 in accordance with coherent carrier recovery techniques, and supplies this downconverted signal to radio frame configurator 228.

The microwave module 246 upconverts to the microwave frequency on the 490 MHz IF output signal generated by the transmit modulator 242, and provides this upconverted signal to a microwave antenna 508 (Figure 12) that transmits the data on link 102. In addition, microwave module 246 receives a microwave frequency signal from link 102, downconverts this signal to a 150 MHz IF signal, and then supplies this downconverted signal to receive demodulator 244.

Figure 4 illustrates a schematic block diagram of the digital signal processing CAM 222 and the radio frame configurator 228 included in the CODEC 220 illustrated in Figure 2. The CAM 222 includes a speed control logic 250 transmission and transmission rate buffers 252. The baud rate control logic 250 receives 100BASE-T 100Mbps Ethernet data packets from the 212 100BASE-T transceiver (Figure 3) through a media independent interface (IIM) between the CAM 222 and the transceiver. 212.

Note that the 100BASE-T data packets are supplied to the transceiver 212 (FIG. 3) as a serial data stream. According to the IEEE 802.3u standard, the serial data stream is encoded using a 4B / 5B scheme. According to the 4B / 5B scheme, each 4-bit (quartet) portion of each 100BASE-T data packet is accompanied by a 1-bit valid data field. Thus, due to the valid data bits, the cable transmission speed for 100BASE-T is actually 125 Mbps even though the serial data communication speed is 100 Mbps, assuming valid data bits are discounted. . Transceiver 212 converts this serial data stream into parallel 4-bit portions of data (nibbles), a valid data signal (RX_DV), and also recovers a clock signal from the data stream. The quartets, the valid data signal, and the clock signal are supplied to the CAM 222 by the transceiver through the IIM interface.

The data nibbles, the valid data signal, and the recovered clock signal are then synchronized to a locally generated clock signal. This locally generated timing signal operates preferably at 27.5 MHz and is obtained from a crystal oscillator at 55 MHz and with an accuracy of 10 parts per million located within the CODEC 220 (figure 3). The baud rate control logic 250 detects each 100BASE-T Ethernet data packet received from transceiver 212. In the preferred embodiment, the baud rate control block 250 then checks each of these 100BASE-T Ethernet data packets for errors using the Frame Check Sequence (SVT) attached to each 100BASE-T Ethernet packet. and extracts from each 100BASE-T Ethernet packet its preamble and frame start delimiter (preferably, the frame check sequence SVT is maintained for each 100BASE-T Ethernet packet). The baud rate control logic 250 also converts each Ethernet data packet from nibbles to bytes.

The baud rate control logic 250 calculates the length of each detected 100BASE-T Ethernet data packet. The baud rate control logic 250 also determines if the packet is too long, too short (a minimum packet), or misaligned.

The baud rate control logic 250 then temporarily stores the packets in the baud rate buffers 252. In the preferred embodiment, the bytes in each packet are timed in the baud rate buffers 252 in accordance with a clock signal recovered from the data. Transfer rate buffers 252 include two FIFO (First-In, First-Out) buffers that have 16K entries, one for packets being transmitted and one for packets being received. Preferably, the FIFO buffers each provide sufficient storage for each entry so that additional information can be stored in the baud rate buffers 252 along with the data byte. Such additional information preferably includes the valid data bit for each nibble and an indication of whether the nibble is payload data or a header for 100BASE-T Ethernet packets. For example, the header may include packet separation codes (eg, an all-zeros byte / octet, with associated non-claimed valid data bits), and packet start codes. Assuming that packet spacing codes are stored, preferably only one packet spacing code representative of the minimum required packet spacing (eg 0.96 μδ) is stored in the baud rate buffers 252.

The baud rate control logic 250 then records the predetermined length of the 100BASE-T Ethernet data packet in a length and status FIFO buffer 254. In addition, the baud rate control logic 250 stores an indication of the status of the packet (eg, if it is too long, too short, or misaligned) in the length and status buffer 254.

Radio frame configurator 228 is coupled to CAM 222 and includes GE / EN block 240 (Figure 3), packet sync / desync block 254, Reed-Solomon encoder / decoder 258 (RS codec), a frame configuration block 260, a pseudo-random number (NP) scrambler / descrambler block 262, a differential encoder / decoder 264, and a constellation mapper 266.

Packet sync / desync block 256 retrieves 100BASE-T Ethernet data packets

ES 2 219 019 T3 stored the transmission rate buffers 252 at a suitable rate that depends, in part, on the data rate used to send data over the wireless link 102. In the preferred embodiment, data extraction from the baud rate buffers 252 for an Ethernet packet does not start until the packet has been fully stored. During periods when a complete packet is not found in the baud rate buffers 252, then a packet separation code is replaced by the packet sync / desync block 254.

In the preferred embodiment of the present invention, the packet sync / desync block 256 reshapes the 100BASE-T Ethernet data packets according to a reshaped frame structure 300 for the 100BASE-T Ethernet data packets illustrated in Figure 5. The reshaped frame structure 300 includes a sync pattern field 302, a length field 304, a data field 306, and a frame check sequence (SVT) field 308.

Recall that the baud rate control logic 250 (FIG. 4) extracts from each 100BASE-T Ethernet data packet its preamble and frame start delimiter before storing the packet in baud rate buffers 252. As each packet is retrieved from the baud rate buffers, the packet sync / desync block 256 adds a sync pattern in field 302 and a length value in field 304 to the packet. The length value is retrieved from the length and status buffer 254.

In the preferred embodiment, finite state machines control sync / desync block 256 to allow retrieval of 100Base-T Ethernet packets from baud rate buffers 252 along with the length and status of each one at a suitable frequency to form radio frames 350 (FIG. 6). A store-and-forward technique is applied to the 100BASE-T Ethernet packets passing through the transmit portion of the rate buffers 252. In this manner, the data packets to be transmitted over the wireless link 102 are fully received in the baud rate buffers 252 and stored there before being formed into a radio frame 350. However, a cut-off technique is preferably applied to the 100BASE-T Ethernet data packets passing through the receiving portion of the baud rate buffers 252. In this manner, the data packets received from the wireless link 102 are retransmitted to the transceiver 212 (FIG. 3) as received, without storing the entire data packet in the baud rate buffers 252.

Table 1 shows the particular bit values ​​for sync pattern field 302 and length value field 304 in accordance with the preferred embodiment of the present invention.

TABLE 1

<td colspan="5">Synchronization field 302</td><td colspan="3">Field 304 of packet length</td><td></td>

<td>Octet 1</td><td>Octet two</td><td>Octet 3</td><td>Octet 4</td><td>Octet 5</td><td>Octet 1</td><td>Octet two</td><td>Octet 3</td><td>Bit</td>

<td> 1</td><td> 1</td><td> 0</td><td> 1</td><td> 0</td><td>G [11]</td><td>G [7]</td><td>G [3]</td><td> 7</td>

<td> 1</td><td> 1</td><td> 0</td><td> 1</td><td> 0</td><td>G [10]</td><td>G [6]</td><td>G (2]</td><td> 6</td>

<td> 0</td><td> 0</td><td> 1</td><td> 0</td><td> 1</td><td>G [9]</td><td>G [5]</td><td>G [1]</td><td> 5</td>

<td> 1</td><td> 1</td><td> 0</td><td> 1</td><td> 0</td><td>G [8]</td><td>G [4]</td><td>G [0]</td><td> 4</td>

<td> 0</td><td> 0</td><td> 1</td><td> 0</td><td> 1</td><td> 0</td><td>L [7]</td><td>L [3]</td><td> 3</td>

<td> 1</td><td> 1</td><td> 0</td><td> 1</td><td> 0</td><td>L [10]</td><td>L [6]</td><td>L [2]</td><td> 2</td>

<td> 1</td><td> 1</td><td> 0</td><td> 1</td><td> 0</td><td>L [9]</td><td>L [5]</td><td>L [1]</td><td> 1</td>

<td> 0</td><td> 0</td><td> 1</td><td> 0</td><td> 1</td><td>L [8]</td><td>....... L [4] ..........</td><td>UQ]</td><td> 0</td>

As shown in Table 1, the sync pattern located in sync pattern field 302 is preferably a five-byte (five-byte) pattern defined by a five-bit Willard code [11010]. Essentially, the Willard code is repeated for each octet, but it is reversed for two of the five octets. The length value placed in length field 304 is preferably an eleven-bit L [10: 0] value that specifies the number of octets (bytes) of payload data contained in data field 306. Thus, the length value L [10: 0] may vary for each packet depending on the length of the data payload included in the 100BASE-T Ethernet packet. In the preferred embodiment, a twelve-bit Golay checksum G [11: 0] is stored for the length value along with the length value in the length field 304, as shown in Table 1. Since field 304 in length is preferably three octets (three bytes), a value of zero (0) is used as a field marker between the length value L [10: 0] and the check sum G [11: 0] Golay.

Referring to Figure 5, the data payload from the Ethernet packet is stored in the field

ES 2 219 019 T3

306 of data. Note that 100BASE-T Ethernet data packets are conventionally of variable length. Specifically, the data payload portion for a conventional 100BASE-T Ethernet packet can range from 64 to 1518 octets (bytes). Thus, the length of the data field 304 can vary between 64 and 1518 bytes.

An important aspect of reforming the Ethernet data packets in the reformed frame 300 is the omission of the 1-bit data field valid for each nibble of the Ethernet packet. Rather, the nibbles are placed contiguously in data field 306. This omission of the valid data bits results in a significant saving in the bandwidth required to transmit the reformed packet frame 300 over the wireless link 102 as compared to also transmitting valid data bits over the wireless link 102. The SVT sequence is held by each Ethernet packet and placed in SVT field 308.

The packet synchronization / desynchronization block 256 also receives data for link control from the GE / EN 240 and to combine this data for link control with the reshaped packet frames 300 to be communicated on the link 102.

RS codec 250 receives reshaped data packet frames 300 and link control commands from packet sync / desync block 256, and performs ReedSolomon (RS) encoding for forwarding error correction. The RS encoded data is then supplied to the frame configuration block 260 where the RS encoded data is formatted according to the radio frames 350 (FIG. 6).

Figure 6 illustrates a radio frame 350 in accordance with the present invention. Radio frame 350 includes a synchronization field 352 for synchronizing a receiver to radio frame 350, an auxiliary field 354 for traffic for network management and link control, which is received from the GE / EN 240, to be communicated on the auxiliary channel 102B of the wireless link 102, a data field 356, and an RS parity field 358. The value placed in the sync field is preferably 47 hexadecimal.

In the preferred embodiment, radio frames 350 are formed and transmitted continuously over wireless link 102 whether data from a complete Ethernet packet is put on hold in baud rate buffers 252 or not ( Figure 4) to be placed in frames 400 of reformed packets. During periods when packet reshaped frames are not available, the data field 356 of the current radio frame 350 is loaded with the padding code (all zeros). Similarly, during periods when there are no waiting orders for the network management to communicate through the auxiliary channel 102B, then the auxiliary field 354 is loaded with padding code (all zeros).

Recall that, according to the preferred embodiment of the invention, the reshaped packet frames 300 have a variable length. However, the data field 356 of each data frame 350 has, according to the preferred embodiment of the present invention, a fixed length. Consequently, the RS encoded data from the RS codec 258 is placed contiguously in the data field 356 of each radio frame 350 so that the boundaries of the reshaped data frame 300 do not have a predetermined relationship to the boundaries of the radio frame 350. For example, a reshaped data frame 300 may span multiple radio frames 350. Alternatively, up to three smaller complete reshaped data frames 300 may be included in a single radio frame 350. In addition, during idle periods between communication of reshaped packets, a padding code is preferably transmitted as a field marker within the data field 356 of each radio frame 350 to satisfy the timing requirements necessary to synchronize the data packets. 100BASE-T Ethernet.

As the radio frames 350 are formed, multiples of the radio frames 350 are combined to form a radio "super frame" 380 (FIG. 7). Figure 7 illustrates a radio superframe 380 in accordance with the present invention. In the preferred embodiment, each radio superframe 380 includes 16 consecutive radio frames 350 (FIG. 6). For the first radio frame 382 of super frame 380, the value placed in sync field 352 is inverted (changed to B8 hexadecimal). In radio frames 384 ranging from the second to the sixteenth, however, the value placed in the sync field 352 remains unchanged. The value placed in the sync field 352 of the first radio frame 386 is also inverted for a next radio super frame 388. This reversal of the timing value for the first radio frame 350 of each radio super frame 380 allows the radio super frames 388 to be detected after reception.

Radio superframe 380 is supplied to NP scrambler / descrambler 262. The NP scrambler / descrambler 262 performs quadrature amplitude modulation (QAM) scrambling throughout the radio superframe 380 except for the inverted sync values ​​placed in the first sync field 352 of each superframe 380. By disabling the NP scrambler / descrambler 262 for inverted sync values, the scrambled superframe 380 can be detected upon receipt. In the preferred embodiment, the scrambling operation assigns each octet (byte) of radio superframe 380 (other than the inverted sync values) to two successive four-bit symbols using a thirteenth order polynomial, as shown by the schematic block diagram of the NP scrambler / descrambler 262 in accordance with the preferred embodiment of the present invention.

Referring to Figure 8, each octet of radio superframe 380 (other than inverted sync values) is divided into two successive four-bit B [3: 0] portions that are applied to the corresponding inputs.

ES 2 219 019 T3 named, illustrated in figure 8. These inputs correspond to components I1, I0, Q1, Q0 of symbols (I&amp;Q) in phase and in quadrature. A return shift register 400 generates the specified thirteenth-order polynomial. The contents of memory cells selected from the return shift register 400 are appended modulo 2 by exclusive-OR logic blocks 402, 404, 406, and 408 with each four-bit b [3: 0] portion of the radio frame. The outputs of exclusive-OR blocks 402, 404, 405, and 408 form I &amp; Q symbol components I1 ', I0', Q1 ', Q0'.

Symbol components I1 ', I0', Q1 ', Q0' are applied to differential encoder / decoder block 264 (FIG. 4). Figure 9 illustrates a schematic block diagram of a differential encoder 264A included in differential encoder / decoder block 264 (Figure 4), and characteristic operations in accordance with the present invention. Encoder 264A forms signal components I1 ", I0", Q1 ", Q0". In the preferred embodiment, encoder 264A is implemented by an appropriately preconditioned lookup table.

The differential encoder encodes the scrambled symbols from the NP scrambler / descrambler 262 in such a way that differentiation of the quantum phase of the transmitted symbols according to the π / 2 modulus recovers the original uncoded data, regardless of which of the four possible alignments quantum phase is selected in decoder 264B illustrated in FIG. 10.

Figure 10 illustrates a schematic block diagram of differential decoder 264B included in differential encoder / decoder 264 (Figure 4) and characteristic equations in accordance with the present invention. In the preferred embodiment, the differential encoder 264B is implemented by a conveniently preconditioned look-up table.

Symbol components I1 ", I0", Q1 ", Q0" formed by encoder 264A are applied to constellation mapper 266 (FIG. 4). Constellation mapper 266 maps four-bit portions of radio frame 350 to sixteen different symbols, as shown in FIG. 11, according to quadrature amplitude modulation (16 QAM) techniques.

FIG. 11 illustrates an assignment constellation for constellation mapper 266 (FIG. 4) in accordance with the present invention. In the preferred embodiment, this constellation is defined by a standard adopted by the Digital Audio Visual Counsel (DAVIC). Input symbol components I1 ", I0", Q1 ", Q0" are mapped to output symbol components Is, Im, Qs, Qm, as shown in Table 2. The assigned symbols are then supplied by constellation mapper 266 (FIG. 4) to transmit modulator 242 (FIG. 3).

TABLE 2

<td>I1 ”, I0”, Q1 ”, Q0” (input)</td><td>Is, Im, Qs, Qm (output)</td>

<td> 0000</td><td> 1010</td>

<td> 0001</td><td> 1110</td>

<td> 0010</td><td> 1001</td>

<td> 0011</td><td> 1000</td>

<td> 0100</td><td> 1011</td>

<td> 0101</td><td> 1111</td>

<td> 0110</td><td> 1101</td>

<td> 0111</td><td> 1100</td>

<td> 1000</td><td> 0110</td>

<td> 1001</td><td> 0111</td>

<td> 1010</td><td> 0101</td>

<td> 1011</td><td> 0001</td>

<td> 1100</td><td> 0010</td>

ES 2 219 019 T3

TABLE 2 (continued)

<td>I1 ”, I0”, Q1 ”, Q0” (input)</td><td>Is, Im, Qs, Qm (output)</td>

<td> 1101</td><td> 0011</td>

<td> 1110</td><td> 0100</td>

<td> 1111</td><td> 0000</td>

The received radio superframes 380 (FIG. 7) are supplied to the constellation mapper 266 (FIG. 4) from the receive demodulator 244 (FIG. 3). During reception of radio superframes 380, each radio superframe 380 is converted back from the symbols Is, Im, Qs, Qm to the symbol components I1 ", I0", Q1 ", Q0" by the constellation mapper 266, which performs a reverse assignment operation based on the relationships shown in Table 2.

In the preferred embodiment of the present invention, the QAM format can be dynamically altered under the control of microprocessor 230 by considering rain fading or interference detected by bit error rates (TEB) or by receiving a command for link control. . For example, the QAM format can be dynamically altered from 16 QAM to 4 QAM. Alternatively, the QAM format can be changed from 16 QAM to 4 QAM and with the application of spectrum spread. Accordingly, the bit rate of the data transmission is reduced, however, the error rate would also be expected to be reduced. Conversely, the QAM format can be dynamically altered from 16 QAM to 64 QAM, resulting in a higher bit rate of data transmission.

Next, the differential decoder 264B (FIG. 10) decodes the symbol components I1 ", I0", Q1 ", Q0" into the symbol components I1 ', I0', Q1 ', Q0'. Next, the radio super frame 380 is detected by the inverted sync values ​​for the first radio frame of each radio super frame 380. The symbol components I1 ', I0', Q1 ', Q0' are then provided to the NP scrambler / descrambler 262 (Figure 4), which converts them back to the original two successive b [3: 0] parts of four bits for each octet of each radio frame 350 (FIG. 6) of the radio super frame 380 (FIG. 7).

Radio frame 350 is then synchronized to radio super frame 380 by detecting the non-inverted sync value in field 352 (FIG. 6) for each radio frame 350. The 258 RS codec (Figure 4) performs forward error correction. For each radio frame 350 that has an error that is uncorrectable by the 258 RS codec, the 258 RS codec provides an indication, preferably by setting a flag, which is stored in the baud rate buffers 252 along with the packet data. in question. For each Ethernet packet formed by the transmission speed control logic 250 that is affected by said uncorrected error, as indicated by the 258 RS codec (figure 4), the transmission error signal TX_ER supplied is used. to the transceiver 212 (Figure 3) through the IIM interface. A link layer response can then be applied to cause the packet to be forwarded.

The reshaped data frames 300 are then passed from the RS codec to the packet sync / desync block 256. In the packet synchronization / desynchronization block 256, the reshaped data frames 300 (FIG. 5), as well as the data for link control and network management, are detected and extracted from the frame 350 structure of radio. For the reshaped frames 300 of data, this is accomplished by a window search technique using a matched filter correlation. The search technique is used to locate the five-octet sync value in sync field 302 (considering the Willard code) for each reformed data frame 300 . When packet synchronization is maintained, the search window preferably encompasses only spacing periods between packets (when data field 356 of radio frame 350 contains the padding code). However, during periods when packet synchronization is not detected, the search window is expanded to encompass the entire packet. Once synchronization is obtained, the window narrows again.

The correlation search is performed by the packet synchronization / desynchronization block 256 using a matched filter that performs a correlation on an octet-by-octet basis. Accumulation by addition runs on 40 data bits at a time (5 bytes), depending on the octets passing through the matched filter. The accumulated value is compared to a predetermined threshold for each octet. When the threshold is exceeded, the beginning of a reshaped data frame 300 is indicated.

Once a sync value is detected, the length value for the packet and the Golay code are read from the length field 304. The length value is verified using the Golay code. If necessary, the length value would be corrected using the Golay code. However, if the length value is corrupted and uncorrectable, the packet is ignored while the search for a next sync value continues.

Assuming the length value is correct or correctable, the reshaped data frame 300 is loaded into the

ES 2 219 019 T3 transmission rate buffers 252 by packet synchronization / desynchronization block 256 in eight-bit (byte) parts for processing in a 100BASE-T Ethernet packet. From the length value, the valid data bit for each byte is also regenerated and stored in the baud rate buffers 252. A single packet separation code is stored in the baud rate buffers 252 to separate each packet. Data for link control and network management from auxiliary field 354 of each received radio frame 350 is provided to microprocessor 230 (FIG. 3) through time division demultiplexing.

The search for a next sync value is then disabled until the end of the reshaped data frame 300, as indicated by the correct or corrected length value.

Reshaped frames 300 of data are retrieved from the packet buffer 252 under the control of the baud rate control logic 250, and returned to the standard 100BASE-T Ethernet format for the IIM interface with the transceiver 212 (Figure 3 ). This is accomplished by resetting the preamble and start-of-frame delimiter for each 100BASE-T Ethernet packet. The conventional 100BASE-T Ethernet packets are then provided to 100BASE-T transceiver 212 (FIG. 3) at an appropriate transmission rate for 100BASE-T transceiver 212. Next, the 212 100BASE-T transceiver transfers the packages to the UPS (Figures 1 and 3). In the preferred embodiment, the baud rate control logic 250 includes a finite state machine to perform the function of retrieving the Ethernet packets from the baud rate buffers 252 and supplying them to the 100BASE-T transceiver 212. Therefore, the transmission rate control logic 250 synchronizes the packets based on a clock signal used for the transfer of the 100BASE-T data packets with the locally generated clock signal, which is used to form and transfer frames. 350 radius (figure 6).

Referring to Figures 3 and 4, in the preferred embodiment, the transmit modulator 242 receives four-bit symbols from the constellation mapper 266 of the radio frame configurator 288 in the CODEC 220 at 27.5 Mbaud. Each symbol is converted to a complex in-phase and quadrature (I&amp;Q) voltage and then pulsed using a raised cosine root filter in transmit modulator 242. Finally, the symbol modulates an intermediate frequency (IF) output signal of 490 MHz. The output level of the signal formed by the transmit modulator 242 is selectively adjustable in a continuous range under the control of the microprocessor 230. Preferably, output level adjustments are made in response to detected rain fade, detected interference, or in response to a command to control the link. The modulated IF signal formed by transmit modulator 242 is supplied to microwave module 246.

The receive demodulator 244 preferably includes a 0-dB / 20-dB IF attenuator in the receive path, which can be selected under the control of the microprocessor 230 depending on the range of action of the link 102. Typically, this attenuator is set for 0 dB. However, for link action ranges of less than about 50 meters, the attenuator is preferably set for 20 dB attenuation. The receive demodulator 244 performs adaptive slope equalization to minimize the effects of distortion caused by transmission on link 102. Additionally, preferably the receive demodulator 244 also includes an adaptive time domain equalizer to perform synchronization of the signals. symbols, and a raised cosine root matched filter process is applied to minimize inter-symbol interference.

Figure 12 illustrates a schematic block diagram of a radio frame Ethernet sync portion 268 of the baud rate control logic 250 (Fig. 4) and a transmit buffer 252A in accordance with the present invention. The transmission buffer 252A forms a part of the transmission rate buffers 252 (FIG. 4). Fast Ethernet 100BASE-T packets and an RXDV signal of valid receive data from the transceiver 212 are received in the transmit buffer 252A, as explained above with reference to Figure 4. Additionally, a clock signal at 25 MHz is derived from the incoming data packet and is used to time the incoming Ethernet data packets in the transmit buffer 252A.

The receive valid data signal RXDV is supplied to a first input of an arbitration logic block 270. In response to the storage of a complete Ethernet packet in the transmit buffer 252A, as indicated by the valid data signal RXDV, the arbitration logic 270 commands a packet counter 272 to increment a count by one. As Ethernet packets are retrieved from the transmit buffer 252A, a delayed valid data signal is also retrieved from the transmit buffer 252A. This delayed valid data signal is applied to a second input of arbitration logic block 270. In response to the removal of an entire Ethernet data packet from the transmit buffer 252A while being supplied to the sync / desync logic block 256, as indicated by the delayed valid data signal, the delay logic block 282 arbitration instructs packet counter 272 to reduce the count by one. Thus, the packet counter 272 keeps an up-to-date count of full data Ethernet packets in the transmit buffer 252A.

This count is provided by the packet counter 272 to a threshold comparer block 274. The threshold comparer block 274 notifies a read packet status machine 276 when a sufficient number of complete Ethernet packets are stored in the transmit buffer 252A to initiate retrieval of the packets from the transmit buffer 252A. In the preferred embodiment, it is only

ES 2 219 019 T3 necessary to store a complete Ethernet packet in the transmit buffer 252A to activate the read packet status machine 276 to retrieve the packet. Once activated to retrieve a packet, the read packet status machine 276 activates a first input to a logic AND gate 278. A second input to logic AND gate 278 receives a read frame enable signal from sync / desync logic 256 (FIG. 4). This read frame enable signal is activated when the sync / desync logic 256 is ready to receive the Ethernet packet data for insertion into a radio frame 350 (FIG. 6).

An output of logic AND gate 278 is coupled to a read input of transmit buffer 252A to retrieve the packet from transmit buffer 252A. As it is being retrieved, the packet is fed to sync / desync logic 256.

An important aspect of the Ethernet to radio frame synchronization portion 268 of the baud rate control logic 250 (FIG. 4) is that it synchronizes the reception of Ethernet data packets to a 25 MHz clock signal, which is asynchronous with respect to the locally generated clock signal. Note that the 25 MHz clock signal is derived from the incoming Ethernet data packets and applied to the transmit buffer 252A to store the packet data while the locally generated clock signal is applied to the transmit buffer 252A. to retrieve Ethernet packet data from the transmit buffer. Thus, the arbitration logic, the packet counter 272 and the threshold comparison logic 274 work according to the obtained 25 MHz clock, while the read packet status machine 276 and the radio frame configurator 288 (Figure 4 ) operate according to the locally generated clock.

In the preferred embodiment, the locally generated clock signal is 27.5 MHz. Since the locally generated clock signal is at a higher transmission rate than the clock signal obtained from the incoming Ethernet packets, in the absence of the synchronization portion 268 of the transmission rate control logic 250, it would also be It is possible that the transmit buffer 252A was flushed while an Ethernet packet was still being received in the transmit buffer 252A. Thus, the timing portion 268 of the baud rate control logic 250 avoids this potential problem.

Assuming that an adaptive countermeasure was employed that slows down the rate at which radio frames 350 are formed (FIG. 6), this would also reduce the rate at which Ethernet packet data is retrieved from the transmit buffer 252A. Assuming this speed was below 25 MHz (eg 13.75 MHz), then it would not be necessary to store a complete packet in the transmit buffer 252A before the retrieval of that packet begins. In the preferred embodiment, in such circumstances, a shortcut is employed whereby the incoming Ethernet data packet is supplied to the radio frame configurator 288 (FIG. 4) before the entire packet is received in the transmit buffer 252A. .

FIG. 13 illustrates a schematic block diagram of a radio to Ethernet frame synchronization portion 280 of the baud rate control logic 250 (FIG. 4) in accordance with the present invention. The receive buffer 252B forms a part of the transmission rate buffers 252 (FIG. 4). In the receive buffer 252B, Fast Ethernet 100BASE-T packets recovered from radio frames 350 (FIG. 6) and an RXDV signal recovered from valid receive data are received from the packet synchronization / desynchronization block 256, as shown. explained above with reference to Figure 4. The internally generated clock signal at 27.5 MHz is synchronous with the radio frames 350 (FIG. 6) and is used to time incoming Ethernet data packets into the receive buffer 252B. The Ethernet data packets stored in the receive buffer 252B are retrieved and provided to the transceiver 212 (FIG. 3) according to a 25 MHz clock.

If a spread spectrum is not employed for the data communicated over link 102, then the clock signal used to time the data within the receive buffer 252B preferably operates at 27.5 MHz. Since the clock signal used to retrieve data from the receive buffer 252B preferably operates at 25 MHz, there is no chance that the receive buffer 252B will empty while an Ethernet packet is still being received in the buffer 252B. of reception.

However, in the event that a spread spectrum is employed for the data communicated over link 102, the clock signal applied to the receive buffer 252B may operate at a lower frequency (for example, 13.75 MHz ) that is synchronous with the internally generated 27.5 MHz clock signal. In this case it would be possible for the receive buffer 252B to be flushed while an Ethernet packet is still being received in the receive buffer 252B. Thus, the timing portion 280 of the baud rate control logic 250 avoids this potential problem, as explained below.

The received receive valid data signal is provided by the sync / desync block 256 (FIG. 4) to a first input of an arbitration logic block 282 and to a read packet status machine 288. In response to the storage of a complete Ethernet packet in the receive buffer 252B, as indicated by the retrieved valid data signal, the arbitration logic 282 instructs a packet counter 284 to increment a count by one. As Ethernet packets are retrieved from the receive buffer 252B, an RXDV signal of valid data is also retrieved from the buffer 252B

ES 2 219 019 T3 receiving. This RXDV valid data signal is used by transceiver 212 (FIG. 3) and is applied to a second input of arbitration logic block 282. In response to a complete Ethernet data packet being removed from receive buffer 252B and delivered to transceiver 212 (FIG. 3), as indicated by the valid data RXDV signal, arbitration logic block 282 commands the packet counter 284 that reduces the count by one. Thus, the packet counter 284 keeps an up-to-date count of the complete Ethernet data packets in the receive buffer 252B.

This count is provided by the packet counter 284 to a threshold comparison block 286. The threshold comparison block 286 notifies a read packet status machine 288 when a sufficient number of complete Ethernet packets have been stored in the receive buffer 252B to initiate retrieval of the packets from the receive buffer 252B. In the preferred embodiment, only one complete Ethernet packet needs to be stored in the receive buffer 252B to activate the read packet status machine 288 to retrieve the packet. Once activated to retrieve a packet, the read packet status machine 288 activates a first input to a logic AND gate 290. A second input to logic AND gate 290 receives a LAN read clock enable signal from transceiver 212 (FIG. 3). This LAN network read clock enable signal is activated when the transceiver 212 is ready to receive the Ethernet packet data for communication to the UPS 106 (Figure 1).

An output of logic AND gate 290 is coupled to a read input of receive buffer 252B to retrieve the packet from receive buffer 252B. As it is being retrieved, the packet is provided to transceiver 212. Accordingly, this aspect of the present invention prevents the receive buffer 252B from emptying while a packet is being delivered from the receive buffer 252B to the transceiver 212 (FIG. 3).

A first alternative approach to avoid overflow in terminal 100 receive buffer 252B during periods when data is being communicated over wireless link 102 at maximum transmission rates, can be implemented when an Ethernet data source ( for example, a terminal on the network 128 'LAN) is operating at a slightly higher speed than the reference clock used to retrieve data from the receive buffer 252B. This approach includes monitoring the current depth of the receive buffer 252B and, as the amount of storage space occupied increases, then the transmission speed of the Ethernet data source is adjusted high using a voltage controlled oscillator. As the amount of storage space occupied decreases, then the transmission rate of the transceiver 212 is adjusted downward. When the buffer is nearly empty, the baud rate is set to the normal 25 MHz level. Both source and local frequency references should be within 100 parts per million above or below the specified 25 MHz. for the IEEE 802.3 Ethernet standard.

A second alternative approach to avoid overflow in terminal 100 receive buffer 252B during periods in which data is being communicated over wireless link 102 at maximum transmission rates, involves reducing the minimum inter-packet space used to send packets removed from receive buffer 252B. For example, instead of using 12-byte times to represent packet spacing, packet spacing can be represented by 11-byte times. This may result in a violation of the IEEE 802.3 standard for minimum inter-packet spacing, however this result is expected to be more desirable than loss of packet data if overflow occurred in 252B memory. intermediate reception.

A third alternative approach to avoid overflow in terminal 100 receive buffer 252B during periods when data is being communicated over wireless link 102 at maximum transmission rates, is for terminal 100 microprocessor 230 to send a command for control of the link to terminal 100 '. This command for link control provides a pause packet to layer two switch 600 '(however, since layer two switch 600' and associated packet buffers 602 'are not shown because terminal 100' is identical to terminal 100, it will be understood that layer two switch 600 and packet buffers 602 illustrated in FIG. 16 have identical counterparts at terminal 100 ', referred to herein as 600 'and 602'). The pause packet causes the switch 600 'to temporarily store packets in its associated packet buffers 602' instead of forwarding these packets to the receive buffer 252B.

Figure 14 illustrates a schematic block diagram of microwave module (MMO) 246 (Figure 3) and microwave antenna 508 in accordance with the present invention. The MMO module 246 constitutes a wireless transceiver for implementing wireless communication over link 102 (Figure 1). The MMO 246 includes a transmit up-converter (C / E-TX) 500 coupled to receive signals from the transmit modulator 242. The C / A TX 500 up converts 490 MHz IF signals received from the transmit modulator 242 to microwave frequency for transmission on link 102. In the preferred embodiment, the transmit frequency on link 102 may be selected under the control of microprocessor 230 in 12.5 MHz hops across two adjacent microwave bands (eg 38.6-39.2 GHz and 39.3-40.0 GHz).

A transmit power amplifier 502 (A / P TX), coupled to the transmit up converter 500, amplifies the microwave signals provided by the transmit up converter 500 to an appropriate level. In the preferred embodiment, the transmit power amplifier 502 has a compression point

ES 2 219 019 T3 of 1 dB at approximately 17 dBm. The nominal power is preferably set at 11 dBm, however, the transmit power is selectively controllable by the microprocessor 230 in response to detected rain fade, detected interference, or in response to a command to control the link.

A transmit subband filter 504, coupled to the output of transmit power amplifier 502, filters out unwanted frequencies from the microwave signal to be transmitted on link 102. Microwave module 246 includes a diplexer 506 coupled to filter 504 of transmission subband. Diplexer 506 couples microwave module 246 to microwave antenna 508 for full duplex communication over link 102 by microwave module 246. Microwave antenna 508 transmits microwave signals on link 102 and receives microwave signals from link 102.

A microwave signal, received from link 102 by antenna 508, is provided to a receive subband filter 510 through diplexer 506. Receive subband filter 510 filters out unwanted frequencies from the received signal and provides a filtered signal to a receiver amplifier 512. low noise (ABR). The received signal is then down-converted, preferably up to 150 MHz IF, by a receive down-converter 514 (C / R RX). However, it will be apparent that a frequency other than 150 MHz can be selected. An intermediate frequency automatic gain control (CGA IF) circuit 516 adjusts the level of the down-converted signal to a predetermined level. An output formed by the CGA FI 516 of circuit 514 is supplied to the receive demodulator 244.

In accordance with the preferred embodiment of the present invention, a microwave frequency synthesizer 518 included in microwave module 246 is synchronized to a reference signal from the precision crystal oscillator and is digitally controlled by microprocessor 230 (Figure 3) with a 12.5 MHz hopping capacity. Two outputs from the frequency synthesizer 516 are each synchronized to the same crystal oscillator reference signal and are provided to the transmit up-converter 500 and the receive down-converter 514 to perform upconversion and upconversion, respectively. downward.

Figure 15 illustrates a perspective view of the microwave antenna 508 and a housing 550 for the external unit 104 (Figures 1 and 3) in accordance with the present invention. Housing 550 protects UE 104 from environmental conditions, such as rain, snow, and sunlight, that can be experienced on rooftops where UE 104 is normally placed. The housing 550 includes a flange 552 to hold the antenna 508 and cooling fins 554 to dissipate the heat generated by the electrical circuits of the UE 104. A cable 556, which is preferably weather resistant and electrically shielded, extends between the UE 104 and the UPS 106 and the connects electrically (figures 1 and 3). Thus, cable 556 includes each of cables 108, 110, 112, and 114 (Figures 1 and 3).

Figure 16 illustrates a schematic block diagram of an alternative embodiment of digital signal processing CAM 222 'and radio frame configurator 228' in accordance with the present invention. The elements illustrated in Figure 16 that have a one-to-one functional correspondence with the elements illustrated in Figure 4 are given the same reference number, but are distinguished by the fact that the reference number is prime. In one aspect, the arrangement shown in Figure 16 differs from that illustrated in Figure 4 in that a layer two switch 600 and associated packet buffer 602 are added.

According to the embodiment of CAM 222 'illustrated in Figure 16, Ethernet switch 600 is coupled to transceivers 212, 214 (Figure 3) and packet buffers 602. Packet buffers 602 provide temporary storage for packets as they are routed through switch 600. Switch 600 is also coupled to microprocessor 230 via interface 604 and to baud rate control logic 250 'via interface 606. Switch 600 may be a conventional layer two Ethernet network switch having a 100BASE-T port coupled to cable 108 and a 10BASET port coupled to cable 110. In the preferred embodiment, switch 600 also includes a 10BASE-T port that is coupled to microprocessor 230 through interface 604 and a 100BASE-T IIM port that is coupled to baud rate control logic 250 '. via interface 606.

Network management and link control traffic in the form of Ethernet packets received by switch 600 from transceiver 212, transceiver 214, or interface 606, and including the CAM address of microprocessor 230 as a destination address, is directed to the microprocessor 230, through interface 604, by switch 600. Similarly, microprocessor 230 sends Ethernet packets to baud rate control logic 250 'through switch 600 for communication on link 102 and to transceivers 212, 214 through switch 600 for communication with the router or switch 116 (Figure 1).

In the preferred embodiment, switch 600 implements a flow control technique according to IEEE 802.3x. In accordance with the present invention, the flow control technique is selectively initiated by the baud rate control logic 250 'by sending a pause packet to switch 600 through interface 606. Each pause packet includes an indication of how much flow control technique must remain active. In response to receiving the pause packet, switch 600 does not provide packets that are received from transceivers 212, 214 or interface 604 at interface 606. Instead, when the flow control technique is active, switch 600 queues such packets by storing them in packet buffers 602. Preferably, the pause signal can be initiated for several hundred milliseconds, while packets are received from transceivers 212, 214, or interface 604, without loss of either.

ES 2 219 019 T3 packages. When the indicated time elapses, the flow control technique is deactivated. Upon deactivating the flow control technique, the switch 600 retrieves the queued packets from the packet buffers 602 and supplies them to the baud rate control logic 250 'through the interface 606.

Baud rate control logic 250 'sends a pause packet with an indicated activation period in response to a hold control signal received from baud rate buffers 252' via signal line 608 . When activated, the hold signal provided via signal line 608 indicates that baud rate buffers 252 'are nearly full. The indicated activation period, included in the pause packet, is appropriate to allow sufficient data to be removed from the baud rate buffers 252 'and communicated on the link 102 via radio frames 350.

As an example of the operation of CAM 222 ', it is assumed that a rain or interference fading on link 102 is detected by an increase in a measured bit error rate (TEB). In response, microprocessor 230 issues a command for link control that causes the data rate for link 102 to be reduced. As a result of this lower data rate for link 102, radio frames 350 are formed less quickly and thus data is removed from rate buffers 252 'at a lower rate. If the reduced data rate results in the baud rate buffers 252 'nearly full, the baud rate buffers 252' activate the hold signal via signal line 608. In response, the baud rate control logic 250 'sends a pause packet to switch 600. Then, while flow control is active, packets received from transceiver 212, 214 or interface 604 for communication on link 102 are temporarily queued in packet buffers 602. Consequently, the CAM 222 'according to the present invention implements a flow control technique to adapt a current data transmission rate on link 102 to a transmission rate at which Ethernet packets are received by CAM 222' from UPS 106 (Figures 1 and 3) without loss of Ethernet packets.

Additionally, the embodiment of CAM 222 'illustrated in Figure 16 includes an encryption / decryption block 612 coupled between logic 250' for speed control and transmission speed buffers 252 '. Accordingly, for packets to be transmitted on link 102, encryption / decryption block 612 encrypts the Ethernet data packets before temporarily storing the data packets in the baud rate buffers 252 '. Conversely, received Ethernet packets from link 102 are decrypted by encryption / decryption block 612 before being provided to switch 600. A buffer 614 coupled to the encryption / decryption block 612 provides a memory buffer for use during encryption / decryption of Ethernet packets. An encryption start control signal line 610, coupled between encryption / decryption block 612 and state / length buffer 254 ', is used by encryption / decryption block 612 to command buffer 254' status / length that provides an encryption label and sequence number to the packet sync / desync block 256 '. This arrangement, which includes encryption / decryption block 612, provides an advantage over the arrangement illustrated in Figure 4 in that data security is improved.

Figure 17 illustrates a frame structure 700 for reshaped 100BASE-T Ethernet packets, formed by CAM 222 'and radio frame configurator 228', illustrated in Figure 16. When the packet is removed from the baud rate buffers 252 'and reshaped for insertion into a radio frame 350 (Figure 6), the encryption tag and sequence number provided by the status buffer 254' / length (FIG. 16) are appended to the reformed frame 700 of packets in a field 702 of cipher labels and in a field 704 of sequence numbers, respectively. The encryption label indicates an appropriate key box, used to encrypt the data, while the sequence number provides timing information to the terminal receiving the reformatted Ethernet data frame 700 from the wireless link 102. The fields of the reformed packet frame 700, illustrated in FIG. 17, having a one-to-one functional correspondence with those shown in FIG. 5 are given the same prime reference number.

Referring to Figure 16, this arrangement also differs from that illustrated in Figure 4 in that it dispenses with the NP scrambler / descrambler 262 and differential encoder / decoder 264 and instead takes an adaptive countermeasures block 616. its place. Adaptive countermeasures block 616 responds to a baud rate change command issued by microprocessor 230 by changing the baud rate at which data is communicated over wireless link 102. The transmission rate at which data is communicated may respond to a detected increase in the bit error rate due to rain fading or it may be to reduce interference with nearby wireless links, such as to reduce interference between subscriber terminals in a point-to-multipoint network.

Figure 18 illustrates a schematic block diagram of the adaptive countermeasures block 616 in accordance with the present invention. A multiplexer 750 is coupled to frame configuration block 260 '(FIG. 16) to communicate radio superframes 380 (FIG. 7) with bounding block 260'. A first NP scrambler / descrambler 262A ', a second NP scrambler / descrambler 262B', and a first differential encoder / decoder 264A 'are each coupled to receive radio superframes 380 from multiplexer 750 depending on the conditioning of multiplexer 750 by part. of the transmission speed change control signal.

ES 2 219 019 T3

In the preferred embodiment, the NP scramblers / descramblers 262A ', 262B', 262C 'perform scrambling over the radio superframes 380 in an identical manner as the NP scrambler / descrambler 262 illustrated in Figures 4 and 8. The superframes 380 radios scrambled by NP scrambler / descrambler 262A 'are provided to a second differential encoder / decoder 264B'. Differential encoders / decoders 264A ', 264B', and 264C 'preferably perform encoding and decoding in an identical manner as differential encoder / decoder 264 illustrated in Figure 4. Next, radio superframes 380 encoded by the second encoder / decoder 264B 'are supplied to a constellation QAM mapper 266'. Preferably, constellation QAM allocator 266 'performs constellation QAM allocation in an identical manner as constellation QAM allocator 266 illustrated in Figures 4 and 16. A multiplexer 756 is coupled to constellation allocator 266' QAM to communicate superframes 380 encoded with the 244 Rx demodulator (figure 3) and the 242 Tx modulator (figure 3). Thus, when a first path is selected through NP scrambler / descrambler 262A ', second differential encoder / decoder 264B', and constellation QAM mapper 266 ', radio superframes 380 are identically conditioned for transmission. and receiving, as when passing through the NP scrambler / descrambler 262, the differential encoder / decoder 264, and the constellation QAM mapper illustrated in FIG. 4. In the preferred embodiment, the first path conditions the radio superframes 380 according to 16 QAM.

The third differential encoder / decoder 264C 'is coupled to the NP scrambler / descrambler 262B' and a quadrature phase shift constellation (MDFC) mapper 752A. The constellation MDFC mapper 752A maps portions of the radio frame 350 to MDFC symbols according to quadrature phase shift keying (MDFC) techniques. Radio superframes 380 are communicated between the constellation MDFC mapper 752A and multiplexer 756. Thus, when a second path is selected through NP scrambler / descrambler 262B ', differential encoder / decoder 264C', and constellation mapper 752A MDFC, radio superframes 380 are conditioned for transmission and reception according to the format MDFC.

A second constellation mapper 752B MDFC is coupled to differential encoder / decoder 264A 'and to NP scrambler / descrambler 262C'. The constellation mapper 752B DFC maps portions of the radio frame 350 to MDFC symbols according to quadrature phase shift keying (MDFC) techniques, identically to the constellation mapper 752A MDFC. Radio superframes 380 are communicated between constellation allocator 752B DFC and multiplexer 756. Thus, when a third path is selected between differential encoder / decoder 264A ', constellation mapper 752B MDFC, and NP scrambler / descrambler 262C', radio superframes 380 are conditioned for transmission and reception in MDFC format with spectrum broadening. Upon receipt, the radio superframes 380 routed through this third path are de-spread and appropriately decoded for communication with frame shaper block 260 '.

In order for radio superframes 380 to be properly received by a receiving terminal (e.g., terminal 100 illustrated in Figure 1), it is important that the appropriate path is selected through adaptive countermeasures block 616 for each superframe 380 of radio. This can be accomplished by notifying the transmitting terminal 100 to the receiving terminal 100 'of the manner and rate of transmission in which the transmitting terminal 100 is transmitting radio superframes 380.

Figure 19 illustrates a graph of received signal level versus time as a result of rain fading. Referring to Figures 1 and 20 and assuming that terminal 100 is receiving data from terminal 100 'over wireless link 102. When rain falls between terminals 100 and 100 ', the level of the microwave carrier signal received by terminal 100, the received signal level (NSR), falls over time as the rain increases over time. Therefore, depending on the weather conditions, the NSR may eventually drop from a normal level to levels below the threshold set at N1-N8. When the NSR is above the threshold level N1, this represents an insubstantial level of rain fade. However, when the NSR is below the threshold level N8, this represents an extreme level of rain fade. The threshold levels N2-N7 represent progressively increasing levels of rain fading between the extremes represented by N1 and N8. The rate at which the NSR (measured slope) decreases can also vary depending on weather conditions. Similarly, as weather conditions improve, the NSR may return to normal. In response to rain fading, the bit error rate (TEB) tends to increase. Thus, the adaptive countermeasures implemented by the present invention can detect the presence of rain fade by measuring the NSR or TEB.

Additionally, TEB tends to increase in response to interference between nearby wireless links. However, an important difference between rain fade and interference is that, in the case of interference, the NSR can remain at a normal level while TEB increases. Accordingly, the adaptive countermeasures implemented by the present invention can detect the effects of interference by measuring TEB.

Accordingly, in the preferred embodiment, the present invention responds to both the measured NSR and the measured TEB. To simplify the following analysis, an example involves a response to a rain fade detected by measuring the NSR. However, it will be apparent that an identical response can be made by measuring TEB. Thus, in the following analysis, the TEB, instead of the NSR, is compared with the various thresholds described (additionally, the> and <operators are interchanged). Additionally, it will be evident that, with

ES 2 219 019 T3 appropriate modifications, a response can be made for both NSR and TEB.

Figure 20 illustrates a flow chart for implementing countermeasures in accordance with the present invention in response to a measured NSR. In the preferred embodiment, microprocessor 230 (Figure 3) is appropriately programmed to implement the flow diagram illustrated in Figure 20. In a first state 800, terminal 100 is configured to communicate data in 16 QAM. The program flow then transitions from state 800 to state 802. In state 802, a determination is made as to whether the NSR has dropped below the threshold level N1. If the NSR has not dropped below the threshold level N1, then the program flow returns to state 800.

However, if the NSR has fallen below the threshold level N2, then the program flow goes to a state 804. In state 804 a determination is made as to whether the rate at which the NSR is changing exceeds a first slope Z1 predefined. If the rate does not exceed the predefined slope Z1, then the program flow transitions from state 804 to state 806. In state 806, a determination is made as to whether the NSR has fallen below the threshold N4. If the NSR has not dropped below the N4 threshold, then the program flow returns to state 800.

However, if the NSR has fallen below the N4 threshold, then the program flow transitions from state 806 to state 808. If the determination made at state 804 resulted in a determination that the rate did exceed the predefined Z1 slope , then the program flow goes from state 804 to state 808. In state 808, the terminal is configured to transmit data according to the MDFC (no spread spectrum). The program flow then transitions from state 808 to state 810.

In state 810, a determination is made as to whether the NSR is above the threshold N5. If the NSR is above the N5 level, then the program flow transitions from state 810 to state 812. In state 812, a determination is made as to whether the rate at which the NSR is changing exceeds a predefined slope Z2. . If the rhythm exceeds the Z2 slope, then the program flow returns to state 800. If the pace does not exceed the Z2 slope, then the program flow goes from state 812 to state 814.

In state 814, a determination is made as to whether the NSR is greater than the threshold level N1. If it is not, then the program flow returns to state 808. If in state 814 the NSR is above threshold N1, then the program flow returns to state 800.

If in state 810 the NSR is not higher than threshold N5, then the program flow goes to state 816. In state 816, a determination is made as to whether the NSR is above threshold N6. If the NSR is above level N6, then the program flow returns to state 808. If in state 810 the NSR is not higher than threshold 816, then the program flow goes from state 816 to state 818. In state 816, a determination is made as to whether the rate at which the NSR is changing exceeds a predefined slope Z3. If the Z3 slope is not exceeded, then the program flow goes from state 818 to state 820.

In state 820, a determination is made as to whether the NSR is less than the threshold N8. If it is not, then the program flow returns to state 808. If in state 820 the NSR is not below threshold N8, the program flow goes to state 822. Additionally, if state 818 exceeds the pending Z3, the program flow goes to state 822. In state 822, terminal 100 is configured to communicate data according to the MDFC with spread spectrum.

From state 822, the program flow proceeds to state 824. In state 824, a determination is made as to whether the NSR is less than threshold N7. If the NSR is not less than level N7, then the program flow returns to state 822. If in state 824 the NSR is greater than threshold N7, then the program flow goes from state 824 to state 826. In state 824 826, a determination is made as to whether the rate of change of the NSR exceeds a predetermined slope Z4. If so, the program flow returns to state 808. If the Z4 slope is not exceeded in state 826, then the program flow goes to state 828.

In state 828, a determination is made as to whether the NSR is above a threshold 828. If it is, the program flow returns to state 808. If in state 828 the NSR does not exceed threshold 828, then the program flow returns to state 822.

An important aspect of the present invention is that a hysteresis is introduced into the flow chart to change the manner of data communication in states 800, 808 and 822 as a function of NSR. So, for example, to switch from 16 QAM to MDFC, the NSR must fall below N2. However, to change from MDFC to 16 QAM, the NSR must increase above N1, where N1 is greater than N2. This hysteresis reduces the frequency at which the way data is communicated is changed and prevents oscillations from occurring between any two of states 800, 808, and 822.

In a point-to-multipoint MAN network, a single network node communicates radio superframes 380 to a plurality of other nodes. Figure 21 illustrates a point-to-multipoint metropolitan area network divided into sectors having interior and exterior radii in accordance with the present invention. An individual node at a center 900 communicates with a plurality of subscriber nodes, designated "r", located at various radial distances from the center 900 and in different directions (sectors). An important advantage of the present invention, which changes in the way it is

ES 2 219 019 T3 communicate data over a wireless link, it can be used to reduce interference between nodes in the same sector, but at different radial distances from the 900 center.

As an example, suppose that a first subscriber node 902 is located in a sector 904 at a radial distance from the center 900 that is less than 2 km. Suppose that a second subscriber node 906 is also located in sector 904, but at a radial distance from the center 900 greater than 2 km and less than 4 km. If both subscriber nodes 902, 906 communicate with center 900 in the same way, there is a probability that communications destined for node 902 will interfere with communications destined for node 906. However, in the preferred embodiment of the present invention, the adaptive countermeasures block 616 (Figures 14 and 16) of the first subscriber node 902 is conditioned to communicate data in a first way (eg, according to 16 QAM), while that the adaptive countermeasures block 616 of the second subscriber node 906 is conditioned to communicate data in a second way (eg, according to the MDFC). The adaptive countermeasures block 616 of the center 900 is conditioned to communicate with any of the nodes 902, 906 by alternately switching between the first and second modes of communication. This is achieved by properly conditioning the baud rate control signal applied to multiplexers 750, 756 (FIG. 18) of hub 900 depending on which node 902, 906 is currently communicating with.

In the preferred embodiment of the present invention, a security authentication protocol is implemented for data security purposes against eavesdropping. Figure 22 illustrates a wireless link 102 between two terminals 100 and 100 'in which an unauthorized terminal 950 is attempting to eavesdrop on a communication between the two terminals 100, 100'. Each terminal 100, 100 'and 950 is preconditioned to periodically authenticate the other terminal on the other side of the communication link. For this purpose, each terminal is assigned a unique password.

Link authentication is carried out as follows: once communication between terminals 100 and 100 'has been established, terminals 100, 100' exchange their passwords. Then, at periodic intervals, terminal 100 sends an interrogation message to terminal 100 '. The interrogation message includes an identification number and a random number. Terminal 100 'receives the random number and calculates a response based on a mathematical combination of the random number and its unique password. The terminal 100 'then sends the calculated response to terminal 100 along with the same identification number that it received.

Terminal 100 then matches the identification number it receives from terminal 100 'with the interrogation message it previously sent and then compares the response it received with an expected response. Terminal 100 'determines the expected response based on its knowledge of the unique password associated with terminal 100' and its knowledge of the random number included in the interrogation. If the response received corresponds to the expected response, terminal 100 'sends a success message to terminal 100'. Then data communication resumes. Each terminal 100, 100 'periodically authenticates the other symmetrically.

However, if the response received does not match the expected response, an alarm is triggered at terminal 100. In response to the alarm, terminal 100 maintains wireless link 102 by sending and receiving radio frames 350 (Figure 6) with terminal 100, however, radio frames 350 sent by terminal 100 no longer carry 100BASE-Ethernet data. T. Instead, the packet separation code is sent. Additionally, terminal 100 is configured to no longer detect and separate 100BASE-T Ethernet packets from received radio frames. Thus, 100BASE-T traffic is interrupted in both directions. The terminals keep trying to re-authenticate the link and, if successful, the 100BASE-T packet communication resumes.

It is important to note that each terminal 100, 100 ', 950 is configured to successfully receive radio frames at all times, but is configured to successfully receive 100BASE-T packet data only if it receives a response to an interrogation message that is corresponds to an expected response. Determining whether a response to an interrogation message is appropriate depends on knowledge of the random number included in the interrogation message.

Suppose that, once wireless link 102 is established, terminal 950 attempts to eavesdrop. This is an unauthorized intruder who is trying to receive data from the link. It is expected that in such a situation, terminal 950 will have its transmitter muted in an attempt to avoid detection. Since the terminal 950's transmitter is muted, it cannot authenticate with any of the terminals 100, 100 '. Thus, although the terminal may receive responses to interrogation messages sent by the terminals 100, 100 ', it cannot equate such a response with an expected response because the terminal 950 will not be aware of the random number sent with the response. Thus, an alarm will be triggered at terminal 950. Once this occurs, terminal 950 can no longer receive 100BASE-T packet data. Consequently, eavesdropping is prevented and data security is maintained.

Figure 23 illustrates an embodiment according to the present invention having multiple digital processing CAMs 222A ", 222B" multiplexed to a single frame configurator 228 ". The CAMs 222A ", 222B" may each be identical to the CAM 222 'illustrated in Figure 16, while the frame configurator 228 "may be identical to the frame configurator 228' illustrated in Figure 16. This embodiment allows multiple 100BASE-T Ethernet packets to be received simultaneously, one for each CAM 222A ", 222B". Ethernet packets are stored tem20

ES 2 219 019 T3 porally on each CAM 222A ", 222B" and then provided to the configurator 228 "through a multiplexer 980 according to time division multiplexing. The time division multiplexed data is then communicated over the wireless link 102. According to this embodiment, the wireless link 102 is configured to communicate data at 200 Mbps. It will be apparent that a number, n, of CAMs can be coupled to the multiplexer 980, thereby achieving a data rate of nx 100 Mbps for the wireless link 102. Such an arrangement is limited by the maximum bandwidth capacity for the wireless link 102.

The present invention has been described in terms of specific embodiments incorporating details to facilitate an understanding of the principles of construction and operation of the invention. Such reference to specific embodiments and details thereof is not intended herein to limit the scope of the claims appended hereto. It will be apparent to those skilled in the art that modifications can be made to the embodiment chosen for illustration without departing from the scope of the invention. Specifically, it will be apparent to one skilled in the art that the device of the present invention could be implemented in several different ways and that the apparatus described above is only illustrative of the preferred embodiment of the invention and is not limiting in any way.