Method and device for reducing the processing time of data in communication networks

[0001] [Technical field to which the invention belongs] The present invention is a packet-oriented application that communicates through a communication network, especially through a mobile communication network such as the Global Mobile Communication System (GSM), Universal Mobile Telephone System (UMTS), General Line Radio Service (GPRS) and an IP network. The present invention relates to an apparatus and a method for improving the processing speed of received data. [0002] [Previous technology] Since the protocol communicates in common, it is defined as the whole of all arrangements with the other party. Therefore, a common binding protocol is essential for exchanging data between two communication network nodes. Protocols are required to be universal and consistently defined, and to be able to connect and communicate different networks beyond the limits of the system on the same basis. [0003] The entire communication protocol in a standard configuration is divided into layers. Each layer solves the challenges assigned by its own protocol. Communication between adjacent layers is guaranteed by a well-defined interface. In this case, the nth layer is directly connected to the n + 1 layer immediately above to provide the service, and is directly connected to the n-1th layer immediately below to receive the service. In addition, it communicates with the service layer n of all lower layers to provide services. Therefore, the logical data flow of the protocol data unit PDU is realized in each protocol layer. On the receiving side, the data is processed in the reverse order. That is, the data is sent directly from the lower layer to the protocol layer immediately above. [0004] The structure of the protocol group may differ depending on the physical network and application. However, there is a limit to guaranteeing communication compatibility between different networks. The protocol group for standardized Internet applications is TCP / IP (Communication Control Protocol / Internet Protocol). It has four layers, the top layer-the application layer-with the application protocol. A communication protocol, for example what is called TCP (Communication Control Protocol), is placed just below it. The Internet Protocol-called IP-forms the network layer. The bottom two layers-the link layer and the physical layer-can be linked to form a network-oriented layer, which are specifically defined depending on the network underneath. The standard structure of the TCP / IP protocol group and the communication link between each layer are shown in Fig. 2. [0005] The TCP transmission protocol provides a reliable transmission service for the flow of bits. Reliability here means error-free, sequence maintenance, and protection against data loss and duplication. Error correction is done by ARQ (Automatic Repeat Request). A copy of the packet for transmission is made by the sender and stored until the transmission of the data packet is reliably approved by the other party. The receiver inspects the received packet, positively confirms that it was received correctly, and deletes the packet received by mistake. In this regard, it should be noted that TCP does not allow the transmission of negative receipt confirmations. The sender concludes that the retransmission of an inaccurate transmitted packet is carried out by a positive approval mechanism, i.e., without positive approval, the packet was not received, depending on the situation. [0006] The bit flow transmitted from the application layer to the TCP layer is segmented by TCP transmitted as an IP datagram. IP datagrams are presented as data packets formatted according to the rules of the IP protocol. It is a characteristic of a diagram that the characteristics of a datagram by exchanging data using a datagram are unreliable. IP does not guarantee that the packet was really transmitted to the receiver. IP datagrams are also out of order or arrive at the receiver twice. Within the limits of this concept, however, it is TCP's job to detect false transmissions and correct any errors that occur. [0007] IP datagrams are also transmitted by the hierarchical principle of the link layer placed directly beneath them. The layer receives and organizes IP datagrams, which are called frames. This is done in a way known as framing. That is, the link layer is packaged as an IP datagram in one or more frames. The frame is separated by a special combination of bits. It is specified what kind of bit sequence the bit combination corresponds to, which is called the separation code at the beginning of the frame, the initial code, and the separation code at the end, and the terminal code. [0008] Apart from the frame, the link layer does two special tasks. The link layer is reliable for error detection. Therefore, erroneously transmitted frames are usually deleted by the receiving link layer. For this purpose, the data packet provides a place where the cyclic code, frame check procedure FCS, cyclic redundancy check CRC, etc. can be applied. This idea is for interpreting multiple data packets. The transmitting side supplements the data packet by receiving the remaining 0s received by the receiving side by a division called a generation polynomial. In this way, error detection is realized. The link layer arbitrarily collects errors by using, for example, the ARQ method. [0009] The link layer protocol is applied directly between physically adjacent network nodes. Many alternative protocols are defined for this purpose. Which protocol is applied between the two network nodes depends on the network connecting the two network nodes. The well-known point-to-point protocol, PPP, is an example of a link layer protocol. PPP performs the first two tasks of the link layer: configuration and error detection. Therefore, there is a special way that PPP works in a numbering mode called RFC1663, but this is usually not used. [0010] In fact, PPP does not support remediation through retransmitted packets, or this method is inefficient if the transmission error rate is high, so special additional protocols are available for data transmission, especially in networks with high error rates. Applies. For example, mobile communication networks are known as networks with high transmission error ratios. GSM (Wide Area System for Mobile Communications) and GPRS (General Packet Radio Service) fall into this category. A special protocol-called RLP (Wireless Linkage Protocol) is applied to the link layer of GSM networks. The bit flow of the RLP segment is received from the PPP layer within the frame, which is usually smaller than a PPP level frame. Error correction is handled by the ARQ method on the basis of the frame. The ARQ function requires consecutively numbered frames. Therefore, each frame receives a series of sequence numbers during classification. In today's practice, bit streams are reliably partitioned and packaged within RLP frames. Thus, as far as the type of data, control data or real data is concerned, the type of data remains unconsidered. Only the RLP layer can know the bit flow. Thereby, the data from two different PPP frames are combined into one RLP frame. The RLP frame then receives the terminal code of the first PPP frame, and also the initial code of the next PPP packet. A solution to this problem is provided in European patent application EP98 113 212.9, which proposes to inspect the flow of bits to the transmitting separator. In this way, when the sender packages the flow of bytes to the RLP packet, it prevents different PPP packets from being distinguished, which prevents the data from the two PPP packets from being combined in the RLP. There is. [0011] Since both protocols, RLP and RLC, are similar to HDLC (High Level Data Link Control Protocol) ISO87, similar functionality is achieved with the RLC protocol in GPRS networks. Differences between protocols exist in the frame creation method. [0012] The purpose of the hierarchy is to ensure that the protocol layer and, above all, the protocols are horizontally independent of each other. In this way, different applications and different communication protocols can communicate via the same network protocol, such as the IP protocol. In addition, the IP protocol can work on different platforms. Therefore, IP diagrams can be transmitted via different physical networks such as GSM and GPRS. [0013] For users, communication at the protocol level is barely visible. Users expect useful systems to support different application services such as email and web browsers. Data often attempts to send packages that are larger than the limits that can be sent over physical links. Therefore, the message is divided into small packets that are arranged continuously. Data partitioning is part of the format. Data formatting is done at each protocol layer. A particular protocol layer, such as the RLP layer, divides the data, i.e. the data is subdivided into smaller blocks of data. Data blocks have different names in different layers. They are called datagrams on the IP protocol layer, frames on the link layer. In addition, data blocks that are not related to each protocol layer are specified by data packets. [0014] The format of the data has control data specifically characterized at each protocol layer. In most cases, control data appends a form called a header at the beginning of a data packet, and / or a form called a tail at the end. The actual data is contained in the user data area. The mechanism is described in detail in the TCP / IP protocol stack below. [0015] According to Figure 3, user data is segmented at the application layer, and control information is added to each data packet. As a result, the data packet is transferred to the TCP transmission layer. The layer adds control data in its header format. The data is transmitted to the network layer. That is, for example, the IP contains appropriate control data such as processing procedure information. An IP diagram is formed in this way and is transmitted to the link layer in the following way. Link-layer protocols, such as PPP, process the received data by adding their own control information, such as separators. The data packet generated at this stage is called a frame. The frame is transmitted via an appropriate network. Data packets arrive at the appropriate layer on the receiving side in different steps. Receiving side processing of this layer can be done to reproduce the transmitted procedure. This is, for example, TCP or RLP receiver processing, but not IP receiver processing. The mechanism of data packaging of the protocol layer is known as encapsulation. The opposite function is called re-encapsulation and is done on the receiving side. Hereinafter, the data packet working in the numbered mode of the RLP frame, the RLC frame, or the PPP frame is referred to as a general name L2ARQ frame. [0016] User data is sent to the receiver in the form of L2ARQ. At the same time, L2ARQ frames are stored in the buffer on the transmitting side. This may be necessary if the packet is retransmitted. The serial number of the L2ARQ frame determines whether the packet was lost during transmission on the receiving side. If one L2ARQ frame is lost, the L2ARQ frame is started to be retransmitted. The corresponding mechanism causes the sender to receive a message of the error that occurred. The packet with the matching number is taken out of the buffer and retransmitted. If the packet is successfully transmitted to the receiver, it is deleted from the sender's buffer. [0017] This mechanism is called numbering mode. The mode provides accurate service by ensuring that accurate data is transmitted from the transmitting side to the receiving side. There is also what is called unnumbered mode. Since the mode uses ARQ processing, no error correction is performed. Therefore, this mode is an inaccurate transmission. [0018] Retransmission of packets, however, means that the order of packets arriving at the receiver does not match the order in which they were transmitted. [0019] In a procedure based on the present invention, arranging L2ARQ frames in the order in which they were transmitted-provided that the link layer protocol supports ARQ-is the role of the link layer protocol. This means that, for example, a received PLP packet is stored in the receiving buffer until the procedure in which the RLP packet is recreated. This also means that when the frame has completely received the packet, and when the frame has the following procedure, the RLP frame RLP frame is generated directly on top of the layer. However, if one frame is retransmitted due to an error, all subsequent frames already received are buffered until the retransmitted frame is received without error. Only when the RLP packet is placed in a procedure generated by a matching sequence number will it pass through the PPP layer. It is deleted prior to the control information. [0020] The PPP layer receiver confirms the PPP frame. Therefore, the receiver looks for a separator. When the PPP frame is fully recognized, the IP diagram is transmitted to the IP layer, where the IP diagram passes through the received TCP segment on the TCP protocol layer. [0021] [0021] The L2ARQ frame is temporarily stored on the link layer in order to bring the frame in the matching procedure, which increases the processing time. This is especially negative for time-lag sensitive applications. To any extent, long delays reduce the efficiency of data processing. For delay-sensitive applications, it can also cause processing to stop. Moreover, this method requires a large amount of buffer on the protocol layer to communicate with. In particular, on the RLP protocol layer, packets are temporarily buffered on said level until the requested procedure is replayed. Long data storage times increase the time for data processing in a hierarchical protocol structure. [0022] [Problems to be Solved by the Invention] Therefore, it is an object of the present invention to provide a method and an apparatus for more efficiently performing data processing on the receiving side of a packet-oriented application in data transmission. It is an object of the present invention to reduce the required memory space, especially on the receiving side. [0023] [Means for solving problems] According to the present invention, the object is provided by claims 1 and 24. [0024] By transmitting the packet completely generated on the link layer directly to the protocol layer immediately above, it is an advantage that long-term temporary storage is not performed. [0025] For this reason, it also offers the advantage that received data is transmitted to the application layer faster, thereby ensuring more stable operation of delay-sensitive applications. [0026] Another advantage is that the received data is not buffered until a sequence of received data is provided, and even if some data packets may not have been received, the fully generated packets. Is generated directly on the protocol layer, so that the storage capacity required by the protocol layer on the corresponding receiving side can be reduced. [0027] Further advantages of the invention are derived from claims 2, 23, 25. [0028] [Example] The invention will be described below with reference to FIG. 1 and claim 1. [0029] According to FIG. 1, the data packet of the first protocol layer is provided to the transmission side 10 and is transmitted directly to the second protocol layer 20 immediately below. The layer packages the received data in the data packet of the second protocol layer 30. The data packet of the second protocol layer does not contain two different data packets of the first protocol layer. Each data packet in the second protocol layer receives a single sequence number. The data packets of the second protocol layer packaged in this way are sent to the network for which step 40 described above is valid, and as a result, are transmitted through the network 50. The individual data packets of the second protocol layer are received on the receiving side 60. The received second protocol layer data packets are sorted in the order of number 70 and stored in the provided buffer 80. They are in turn checked to recognize the first protocol layer 90 data packets. If a second protocol layer data packet is received, it is checked to see if this data contains the first protocol layer separator. If included, either the initial code or the terminal code of the first protocol layer is important. In the case of the initial supplement, it means that the data packets that follow in the second protocol layer belong to the new data packets in the first protocol layer. The data packets of the second protocol layer are stored in the buffer until the data packets of the first protocol layer are completely received by 100. This is detected by the receipt of the data packet in the second protocol layer, the data area contains a terminal code, and further includes one of the following in the procedure: Only the only fully generated data packet in the first protocol layer is generated directly in the protocol layer above 110. [0030] Below, the claims<u style="single">23</u>The invention according to the above will be described. [0031] The data packets of the first protocol layer on the second protocol layer and their arrangement according to the transmission order are realized by providing the data packets of the first protocol layer to the second protocol layer. The data packet is transmitted through the network provided by the transmission means. The receiving means for receiving the data packet on the receiving side receives the packet. A classification means for the classification of received data packets, which is incorporated into a continuous order and stored in a buffer for temporary storage of received data packets in the second protocol layer. The data packets of the second protocol layer are inspected for recognizability of the data packets of the first protocol layer. This is done by means of detection in the first protocol layer to detect fully coupled data packets. As a result, the data packets fully generated in the first protocol layer are inspected by the inspection means as a combination of data streams. Therefore, the inspected data packet is released by the release means to release the fully generated data to the first protocol layer. [0032] An area of application of the present invention is the area of Internet applications through mobile data networks, such as GSM. Possible applications of the present invention have been described in detail by way of examples of the present invention. As a result, the processing of data up to the release of the fully generated data packet on the receiving side is illustrated as an application on the transmitting side. [0033] For this purpose, the network system of FIG. 4 schematically shown is used. Mobile subscriber-to-subscriber communications, such as mobile stations and a collection of subscribers in a fixed network, a server, are schematically illustrated by this. The upper part of the figure shows the physical connection that matches the communication unit, and the lower part constitutes a logical connection of complex protocols. [0034] The mobile station MS is, for example, a laptop computer. The laptop computer connects through a terminal connection function (TAF). The work is, for example, mobile station MS such as mobile phone and PCMCIA (Personal Computer Memory Card International). It is done by Association). The mobile station MS communicates with the BTS (base station radio station), which again communicates with the BSC (base station controller). The connection to the public analog telephone network, called the Public Switched Telephone Network (PTSN), is realized by the modem integrated in the interworking function IWF. The interworking function IWF is called the Mobile Switching Center, Mobile Service Switching Center (MSC). In addition, the connection is made through the public switched telephone network PTSN to an Internet Service Provider (ISP) that has a network transmission node to the Internet. Connections to end subscribers, servers, are established through the Internet. For clarity, connections over the Internet are not shown in more detail in Figure 4. [0035] The application runs independently of the lower protocol layer. The transmission of the data thus generated is performed in a way that is clear to the user. This is also the goal of the hierarchical structure of the protocol stack, i.e., ensuring optimal and stable transmission without including users in the system. It is, however, expected from systems that support all applications used by users, such as video data transmission and access to the Internet. Different applications, however, have different requirements in the system. [0036] Only for certain internet applications such as banking transactions, such as the demand for secure transmission protocols, this approach guarantees error-free data flow during money transactions over the internet. Secure transmission of data is guaranteed by what is called the transmission control protocol TCP. [0037] In contrast, in the case of video transmission, the use of protocols that reward the reliable security of data transmission is not required. Reliable data flow is safe even if there is a possibility of long delay time during transmission. In the case of video transmission, there is a better guarantee for faster transmission of data in the procedure in order to get a realistic feel in the presentation of the video. The errors that may occur during transmission are limited and can be tolerated when broadcasting video. For this reason, error correction protocols are not used in video transmission. An example of such a protocol at the transmission stage is the User Datagram Protocol UDP. [0038] Most of the several applications that users use in a session, for example, whenever a user wants to send an email and transmit video simultaneously in the background. In this case, the user creates two different streams of data. E-mail transmission is based on TCP, and video transmission is based on UDP. Another example is internet access. Multiple Internet pages are open during a session, which are often located on different servers. However, the flow of generated data is exclusively the flow of TCP. Different data flows, in this case, the recipients are different. [0039] The feature considers a network protocol layer, such as the IP layer. The layers are bundled to form a proprietary format packet that receives packets from the transmission protocol layer. Figure 5 illustrates the format of IP packets. The packet contains, for example, data in an IP protocol version of IPv4 or IPv6 version. This should not be shown in detail in 5. In addition, IP data formats are provided in areas that contain information about transmission protocols. In the case of the UDP protocol, the bit binding goes into the region, which corresponds to the UDP designation. [0040] The decisive factor that identifies the flow of data, however, is not only the type of protocol, but also the address contained in the IP header. This is checked in the TCP header with further steps to find differences in data flow if the sender's IP address and the receiver's IP address match the two IP packets shown in Figure 5. Different port numbers are assigned to different data streams. Since communication with the other party is guaranteed for the port numbers, the data flows that match in the transmission stages are the same. The header of a TCP packet contains information about the port number, which is compared when distinguishing the flow of data. When the port numbers on the transmitting side and the receiving side are the same, the data flow is the same. If the addresses are different, i.e. if both IP addresses and port numbers are different from each other, the receivers are different and the data flow is different. The above mechanism is implemented in the version of IP used today, which is called the Internet Protocol version IPv4. The next IP version is 6, IPv6, the definition is basically the same. Here, different data flows are distinguished by what is called data flow identification-or flow identification. The described method can be transferred over IPv6, and each protocol stack is recognizable, basically like a data flow. [0041] Packets on the IP layer Dependencies on whether to obtain the same or different data flows, differences are made between the two modes. It relates to the case of IP packets, which is called the same data flow, intra data flow or intra flow mode. The mode is specified during the interflow mode, which distinguishes IP packets belonging to different data. [0042] According to FIG. 4, an example of data flow from the server to the mobile station MS, for which the coupling on the transmission layer between the mobile station and the server on the Internet has already been established, is shown below. In this example, the communication unit and the communication protocol will be described in more detail. [0043] The IP diagram packaged on the network layer is transmitted via the Internet called the Internet service provider ISP. The ISP transmits the received IP packet to the PPP layer. The layer generates a flow of bits formatted in a PPP frame from the obtained data. Add a separator to distinguish between received packets. As a result, PPP frames are provided for analog transmission. The ISP provides a modem for modulating and transmitting data into an analog signal according to the transmission ratio and mode. In the illustrated example, it is connected or executed via an analog network. PSTN, this is a v32 modem. If connected via an ISDN network, the v.110 protocol is used, for example. V.42 is used to control the flow, that is, to prevent data overflow in the function IWF in the Internet. The above functions are consistent with the functions of the wireless link protocol RPL in GSM. [0044] Within the Internet function IWF, the conversion of received data into a format is done by running GSM. [0045] In this means, the bit flow of the PPP layer is released to the RLP layer. The layer packages the received bit stream into an RLP frame. The format of the RLP frame is shown in Figure 6. The RLP frame has 240 bits. 16 bits are scheduled for header information, 24 bits are the frame check procedure FCS. A decisive factor when packaging a PPP frame within an RLP frame is that higher protocol layer data packets are not directly apparent to the RLP layer in the received bit stream. This means that the RLP layer cannot make a difference between PPP frames, or between IP diagrams, and between packets in the transmission layer. Bit flow must check separators to distinguish packets. This is necessary to avoid packaging data from two different PPP frames into one RLP frame. Each newly generated RLP frame is provided with a sequence number. Packets placed in this way are transmitted over the provided mobile network. During transmission, the RLP frame procedure is confused by the transmission errors that occur and the ARQ processing for the connection. The frame is received by the receiver within the alternate procedure. The receiver checks the received RLP frame against the sequence number in the procedure for finding the position of the RLP frame in. At another stage, check if the RLP frame was received with a separator. If it contains an initial sign, it will be detected as the first frame in subsequent PPP frames and will be stored in the buffer on the matching position. RLP frames that indicate subsequent consecutive numbers are stored in a buffer on the matching position. This continues until the PPP frame receives the state of the fully generated frame. PPP frames are fully generated, if both initial and terminating codes are reliably received, and all RLP frames are reliably received If, and if they are placed in the correct procedure between RLP frames, including the initial and terminal symbols, without any jumps. The frame is re-encapsulated, that is, the control data of the RLP protocol layer is deleted before the RLP frame is buffered. [0046] Not only are PPP packets distinguished when RLP frames are checked, but they can also be extended to detect frame check IP packets. This is the basis for distinguishing between intraflow mode and interflow mode. As already mentioned, the IP header contains information that considers the use of transmission protocols and contains the same address. Due to the fact that the entire IP diagram fits the PPP frame, the check that ends the PPP frame can be specified for the recognition of the IP diagram and the information, i.e. whether the IP diagrams of different data flows are the same. [0047] For this purpose, the frame control data is checked after the complete PPP frame has been generated. The data is specifically inspected against the control data at each protocol layer. It is approved using approval means for approval of fully combined data packets. Information considering the control data of each layer is useful in the above means, the basis of which is the decision made from the control data extracted from the link layer. Especially from PPP, the control data of the IP control layer starts. The format for data re-encapsulation is standardized at each layer. The realization must be made by the mechanism described above, which is similar to the legitimate standard of encapsulation. A more rigorous description of the IP diagram is given in the examples described below. [0048] The IP diagram is transmitted directly over the protocol layer-transmission layer-. TCP packets are numbered equally and consecutively, and a pre-numbered procedure for TCP packets on the transmission layer is generated. In other words, TCP is reliable for the placement of TCP packets in the correct procedure, at this stage inappropriate packets are also detected, and the error is removed by initiating a query for packet retransmissions. [0049] TCP is reliable for producing accurate procedures for transmitted TCP packets. It no longer requires the same behavior as the network protocol layer. This specifically allows IP diagrams received within the modified procedure. The cause for the inaccurate procedure of receiving IP diagrams is asynchronous transmission. Individual packets can take different paths. Packets are sent overtaking each other along their path, thereby reaching the receiver in a strange way. The transmission layer, especially TCP, is reliable for procedure generation. The extent to which the order of IP packets at the network layer changes is not important. This means that the efficiency of packet processing is not particularly affected, even if the order changes later depending on the RLP protocol layer. [0050] Package reordering allowed for the same reason in UDP [0051] Hereinafter, the realization of the invention according to claim 16 with respect to the interflow mode will be described in more detail with reference to FIG. [0052] In interflow mode, packets belonging to different data streams are distinguished. For this purpose, fully generated PPP frames are inspected, as described above. In said mode, PPP frames are already emitted by the RLP receiver, first when they are received completely and accurately, and secondly, PPP frames are included in the data that may be buffered by the RLP receiver. No, it belongs to the same data flow of PPP frames to be released. [0053] After the control data of the IP layer, the recognized transmission protocol area can be searched in the data. Clearly different data flows are involved when the entrance to the region is different in the inspected PPP frame. If, however, there is an entrance that takes into account the transmission protocol, the IP addresses on the transmitting and receiving sides are checked. If the addresses match, the transmitting and receiving port numbers are checked. If there is no difference, it will be detected during this test. PPP frames with the same data flow are relevant. [0054] According to FIG. 7, the following cases are assumed, the transmission side transmits data from two different data streams, UDP data stream and TCP data stream, 170. The PPP data packet is generated from the data in the encapsulation process 180. Two types of PPP data packets, PPP (IP (TCP (n))) and PPP (IP (UDP (n))), are distinct, which are related to UDP data flow or TCP data flow dependencies. Will be done. n specifies the sequence number of the UDP packet or TCP packet by this. According to Figure 7, two UDP packets PPP (IP (UDP (1))), PPP (IP (UDP (2))) and two TCP packets, PPP (IP (TCP (1))), PPP (IP). (TCP (2))) is generated on the PPP protocol layer. They are transmitted to the RLP protocol layer, the package is the same in continuous RLP frames RLP (1), RLP (2), ... RLP (12), 190. It is already mentioned above that the no difference between different data packages in the protocol layer is created on the RLP protocol layer. According to FIG. 7, the data packet PPP (IP (TCP (1))) is divided into RLP (1), RLP (2), RLP (3) and RLP (4). Another data packet in the network protocol layer is split in the same way. The final processing of the RLP frame is to transmit it via network 200. During transmission, changes in the order of RLP frames may occur, due to the common repetition of inaccurate RLP transmission of TCP data flow. [0055] According to Figure 7, the receiver receives the RLP frame RLP (1) first, 210, and then the RLP frames RLP (5), RLP (6), RLP (7), 220. These are recognized as fully received packets. As a result, the packet is inspected in the procedure for detecting the type of data flow. It is recognized as a UDP packet PPP (IP (UDP (1))) and is emitted to PPP layer 230. The PPP layer, however, does not have a PPP frame in the same data stream, probably containing data buffered by the RLP receiver. This example only allows releasing PPP frames to PPP protocol layers that belong to different data streams. Or, for the same data flow, however, only the correct procedure, considering the RLP frame numbering. [0056] According to Figure 7, RLP frames, RLP (8), RLP (9), RLP (10) are received next, 240. These are recognized as complete PPP packet PPP (IP (UDP (2))) 250 and are buffered. The RLP protocol has information on the first UDP packet PPP (IP (UDP (1))) that belongs to the same stream of data that has already been released, and is a PPP (IP (UDP (2)) ))) Determined based on the information. A PPP frame for a TCP packet has not yet been fully generated, a PPP (IP (TCP (1))) containing only RLP (1), which is still kept in the buffer. If, however, PPP (IP (UDP) (1))) is also not perfect, then PPP (IP (UDP (2))) is completely generated, and PPP (IP (IP (UDP (1)))) is generated. Can be stored in the buffer. [0057] The following example proposes an extended implementation, where it accepts fully generated PPP frame discharges that belong to both different and identical data streams. [0058] [0058] Hereinafter, the above embodiment will be described in more detail with reference to FIG. 8 and claim 17. [0059] Temporarily assume a poor transmission quality connection, which occurs during the transmission of high speed PPP packets, first PPP (IP (UDP (2))) is fully received. This happens to receive frames RLP (8), RLP (9), RLP (10), 280. Intermediate memory contains only one RLP frame, RLP (5), 270. All fully generated frames, which also belong to the same flow of data, are released to allow interflow mode. This means exclusively allowing the completeness of the PPP frame. The higher layers then respond to the packet placement in the correct procedure. RLP (1) is first received, and it constitutes the first frame of incompletely generated PPP (IP (TCP (1))), which is maintained in the buffer 260. [0060] As described above, the present invention has been introduced by a typical application in GSM. Other networks have the same potential applications, such as GPRS networks. The network is designed for the transmission of packet-oriented applications from the transmitting side to the receiving side. This protocol structure can be compared in both cases. [0061] The present invention can also be applied to an environment in which only one link protocol is provided. This means that single-link protocols are implemented in GSM instead of PPP and RLP, or in GPRS instead of LLC and RLC. In this case, the protocol is required to work in a reliable mode. It is possible to find this form of realization, for example, in UMTS. [Simple explanation of drawings] Hereinafter, the invention will be described in more detail with reference to examples and figures of the invention. FIG. 1 is a flowchart of the method of the present invention. FIG. 2 is a diagram of the protocol layer of the Internet. FIG. 3 is a schematic diagram of user data. FIG. 4 is a diagram of a network system. FIG. 5 is an internet protocol diagram. FIG. 6 is a diagram of an RLP frame. FIG. 7 is a diagram of an interflow mode. FIG. 8 is a diagram of an intraflow mode.