Method and arrangement for packet-switched data transmission

Method and system for packet-switched data transmission

Background of the invention

The invention relates to packet-switched data transmission and in particular to packet-switched data transmission of a telecommunication network belonging to a digital radiotelephone system when several data transmission speeds are used over the air interface of the radiotelephone system.

In most digital radiotelephone systems, the radio path carrier is shared among multiple users using, for example, the Time Division Multiple Access (TDMA) method or the Code Division Multiple Access (CDMA) method. What these methods have in common is that the call is coded in parts of a certain length into call periods (blocks) transmitted in frames on the radio path. These call periods consist of constant length channel coding information, payload, and error correction information, the length of which is determined by the error correction level used. In particular, several different error correction levels have been defined for data calls. Thus, the length of the payload transmitted on the radio path during a call period depends on the error correction level used in the call and varies from call to call. In general, only a payload is transmitted in a transmission network from a radio call period, which is characterized in that the payloads of the same call are of the same length, but the payloads of two different calls may be of different lengths. A similar situation may exist in other telecommunication systems, where the protocol transmitting data over some interface of the network uses different payloads. In the following, the problems of the prior art will be described with emphasis on aspects of the radio system, without, however, limiting the invention to radio systems.

Packet-switched transmission networks are often based on either ATM (Asynchronous Transfer Mode) or HDLC (High-level Data Link Control).

In the asynchronous mode, the ATM information is transmitted in the standard-sized 53-byte ATM cells of Figure 1A, each cell having five bytes of the cell header HE (Header) and the remaining 48 bytes of actual information IF (Information Field). Cell fields are sent from left to right. To put it simply, the user information to be transmitted in ATM technology is broken into fixed-length snippets, and each snippet is placed in the information field IF of the ATM cell. If the user information is less than 48 bytes, padding bits are sent in the ATM cell. The problem with ATM technology is that the payload to be transmitted on the radio path is usually not 48 bytes, or 384 bits, which means that the transmission capacity is underutilized due to the use of padding bits.

In addition, if the payload of a radio path call period is greater than 48 bytes, the contents of one call period must be transmitted in more than one ATM cell. In this case, in a congested network, the quality of the connection may suffer and the transmission capacity may be underutilized after the radio packet is delayed due to, for example, the last ATM cell in the call period arriving too late at the transmitting network node due to network load. All bits in the call period must be received before the call period to be transmitted on the radio path can be encoded.

Another problem with ATM technology is that adding padding bits and / or dividing the payload into multiple cells causes additional processing of the information to be transmitted when entering and leaving the transmission network.

If one standard-sized cell or packet based on ATM technology is used, the information field length of which is defined as the length of one payload to be transmitted on the radio path, it is possible to reduce the use of padding bits. However, the problem remains that other payload lengths transmitted on the radio path have to use padding bits and / or divide the payload into several different packets.

One packet switching mechanism based on the protocol control procedure HDLC suitable for synchronized transmission is Frame Relay FR, which is utilized, for example, in the General Packet Radio Service (GPRS). GPRS is one of the topics in the phase 2+ standardization work of the pan-European GSM system (Global System for Mobile Communications) at ETSI (European Telecommunications Standard Institute). Figure 1B shows the basic structure of an FR frame according to the HDLC protocol without start / end bits (for asynchronous links) or bits or octets placed in the frame for transparency. The frame has five fields. The fields in the frame are sent in order from left to right. The frame begins and ends with an octet-long flag F (Flag), which is a specified bit sequence starting and ending with a zero bit and containing 6 consecutive first bits between them. The flag starting the frame is followed by the address field AD (Address) and is followed by the information field I (Information). The information field I is not of fixed length, but contains an integer number of octets of data to be transferred and padding bits. The information field is followed by a 2 octet Frame Check Sequence (FCS).

HDLC, and thus also FR, contains the so-called bit stuffing protocol. In this case, the presence of certain bit patterns, such as content reserved for the flag, in the information field is prevented by adding padding bits. However, these padding bits cause problems in the transmission network. In some HDLC-based mechanisms, a payload of a certain size is always transmitted, with the length of the frame being tied to the content of the data to be transmitted. In a network using FR technology5, the parties to the transmission may negotiate the length of the frame to be used before the transmission. If the length of the frame and thus also the length of the information field is constant, the length of the payload transmitted in the frame varies according to the content of the data to be transmitted. In this case, for example, of the successive radio cycles of the same call, the payload of the first may fit in one frame, while the payload of the latter may have to be divided into several frames, even if the payloads on the radio path are equal. If the length of the frame is agreed to be so long that no division is required, padding bits vary in number from frame to frame - depending entirely on the content of the data to be transmitted.

The problem with HDLC and thus also with FR technology is the need for processing caused by padding bits when entering and leaving the network, as well as wasting transmission capacity. In addition, variation in frame length causes variation in transmission times. Similarly, a variation in the length of the payload moving in the frame causes a variation in the payload transfer times. If the transmission times vary too much, the speech quality will suffer. Especially in real-time speech, it is important that transmission times are as constant as possible throughout the speech. In addition, in radiotelephone systems, the variation in the transmission times of data call packets is also detrimental, because the call periods must be transmitted on the radio path at a constant rate according to the radio path protocol, regardless of their content.

Brief description of the invention

It is an object of the present invention to provide a method and a communication system which can solve the above-mentioned problems and maximize the utilization of the transmission capacity of a packet-switched transmission network, as well as minimize the payload processing required by the transmission network. This amount is achieved by a method according to the invention for transmitting data in a telecommunication system comprising at least one packet-switched transmission network, in which data is transmitted in data packets of a constant length, the fixed lengths being defined by a predetermined number. The method according to the invention is characterized in that a constant length code is maintained for each constant length, and a length code corresponding to its length is added to the header of the packet to indicate the length of the packet to be transferred.

The invention further relates to a system in which the method according to the invention can be utilized. The system is a communication system utilizing wireless data transmission, comprising a packet-switched transmission network for transmitting data in fixed-length data packets of a predetermined number of fixed lengths, the transmission network comprising at least a first node for transmitting data packets to the transmission network and a second node for receiving data packets. said first node comprising receiving means for receiving information arriving in the transmission network, editing means for assembling data packets, and transmitting means for transmitting data packets) to the transmission network, and said second node comprising at least receiving means for receiving the data packet. The system is characterized in that the editing means are adapted to include in the data packet a header containing at least a bit sequence length code to indicate the constant length used in the packet, and second node receiving means adapted to identify the length of the received data packet from the packet header length code.

The invention is based on the idea that the length of packets can be expressed in a short bit sequence when using constant length packets with predetermined lengths.

An advantage of the method and system according to the invention is that in the transmission network it is possible to send packets of different lengths without padding bits using a short header. This reduces the redundancy of the data to be transferred. After all, a one-bit length code can indicate two different packet lengths, a two-bit length four, a three-bit eight, etc., regardless of the actual packet length. When the packet sizes to be used are appropriately selected, the packets do not need to be filled with padding bits.

In a preferred embodiment of the invention, the number of packet lengths is defined as many as the net data rates over the air interface. This has the advantage that a separate packet size can be defined for each payload moving over the air interface, and these packet sizes can be used to optimize the use of the transmission network.

In another preferred embodiment of the invention, the lengths of the packets are defined so that the payload transferred therein is equal to the payload transferred over the air interface. This has the advantage that the payload moving over the air35 interface can be placed in a packet without adding full 108902 tebits and dividing the payload into, for example, two smaller packets. This shortens transmission times and improves the utilization of transmission capacity.

In a preferred embodiment of the invention, error checking information is added to the packet according to the error correction level to be transmitted over the air interface. In this application, error checking refers to the detection and possible correction of an error in accordance with the error correction protocol used in the embodiment. This has the advantage that checking and correcting transmission errors in the transmission network corresponds considerably to error correction of the more error-prone air interface. If data is transmitted at the air interface without error correction, it can also be transmitted in the transmission network without error checking.

In another preferred embodiment of the invention, header verification information is added to the header of the data packet, and the forwarding of the received packet is started at the relaying network node after checking the correctness of the header. This has the advantage that the packet can be forwarded before it has been received in its entirety, thus shortening the packet transmission time. Checking the correctness of the header ensures that the packet goes to the correct address. This improves the capacity utilization of the transmission network because packets with an incorrect address are not forwarded.

In a preferred embodiment of the invention, data packets containing user information are separated from control packets on the basis of a bit in the header, and information indicating the length of the packet is added to the control packet. This has the advantage that data and control packets can be transmitted using the same transmission link, and that the lengths of the control packets can vary and the signaling information does not have to be matched to the lengths of the data packets. After all, control packets are rarely sent and are short. This embodiment avoids the use of padding bits. In addition, the embodiment makes it possible to send even very long control packets in the transmission network, for example in connection with updating the network.

Preferred embodiments of the method and system according to the invention appear from the appended dependent claims 2 to 6 and 8 12.

List of figures

The invention will now be described in more detail in connection with preferred embodiments, with reference to the accompanying drawings, in which Figure 1A shows an ATM cell structure, Figure 1B shows a Frame Relay frame structure, Figure 2 shows a block diagram of a TETRA communication system, Figure 3 illustrates 4 shows a frame according to the invention, and Fig. 5 illustrates the content of the header of a frame according to the invention, and Figure 6 shows a block diagram of a network node.

Detailed description of the invention

The invention is described below by way of example, assuming that the telecommunication system is a digital radio system according to the TETRA standard (Trans-European Trunked Radio or Terrestrial Trunked Radio) defined by ETSI (European Telecommunications Standards Institute). The TETRA standard standardizes e.g. interfaces to other networks, air interface and interface to another network size according to the TETRA standard. However, the TETRA standard does not standardize the internal structure of the transmission network, so it is well suited as an example network. However, the invention is not limited to radio systems or other wireless systems, but it will be obvious to a person skilled in the art how the invention can be applied in other telecommunication systems both in wireless data transmission networks and in fixed networks.

Figure 2 shows one example of the structure of the TETRA system 1. The mobile station MS (Mobile Station) is connected to the transmission network 2 via the base station BS by radio Air. The radio interface is defined in the TETRA standard. Each base station BS (Base Station) includes a Node N connected by a communication line to one of the fixed exchange network TETRA exchanges DXT (Digital Exchange for TETRA). The TETRA exchanges DXT are connected in a fixed connection to other exchanges DXT and to the TETRA node exchange DXTc (Digital Central Exchange for TETRA), which is a exchange to which other exchanges DXT and / or other nodes DXTc are connected to provide alternative traffic routes. The interface to another TETRA network in this example is located in the node exchange DXTc, but it can also be located in other exchanges DXT. The external interfaces defined by the standard to the Public Switched Telephone Network (PSTN), the Integrated Services Digital Network (ISDN), the Private Automatic Branch Exchange (PABX) and the Packet Data Network (PDN) are, in this example, located in a single 08902 tea center. DXT, but they can be even in any center. Thus, all nodes N, DXT and DXTc of the transmission network 2 can both send data packets to the transmission network 2 and receive them from the transmission network 2. The operation of the nodes will be described in more detail in connection with Fig. 6 using DXT as an example node. The TETRA transmission network also comprises other interfaces and peripherals not shown in the figure. These include, for example, network management systems and site systems.

Figure 3 illustrates the relationship between radio path call periods and packets transmitted in a transmission network. When transmitting information on the radio10 path, frames are transmitted on a carrier in succession. One frame contains carrier-specific channels, i.e. connections, and thus a call period on the channel belonging to the call intended for the receiving mobile station. When in a TETRA system or other similar system one call is treated as separate speeches, in this description a call is a speech. Correspondingly, the base station receives successive frames in which a channel is allocated for each connection. According to currently used carrier channel allocation methods, the frames are either TDMA or CDMA frames. The transmission time of the frames depends on the radio system used. For example, in a system conforming to the TETRA standard, it takes about 60 ms to transmit one TDMA20 frame, and in a narrowband data transmission implemented by the CDMA method, it takes about 20 ms to transmit a frame.

Figure 3 shows the channels C1, C2, C3 and C4 to be transmitted on a carrier in a first preferred embodiment of the invention according to the TETRA standard. The TETRA compliant radio interface Air is based on a four-channel time division multiplexed carrier with a bandwidth of 25 kHz. In the TETRA standard, a channel is called a time slot. Hereinafter, however, the word channel will be used in the same sense. For the sake of clarity, only one carrier is shown in the figure and only in the other direction.

One base station can have many carriers and there are carriers in both the uplink and downlink directions. One call is usually reserved for each call, but the TETRA standard radio system allows more channels to be reserved for one call. However, for clarity, it is assumed that one call reserves one channel. Referring to Fig. 3, it is assumed that unprotected data is transmitted on terminal C1, whereby the data rate is

7200 bits per second and the actual data content of the call period, i.e. the payload

08902

PL-1 is 432 bits, or 54 bytes. Channel C3 transmits protected data, whereby the data rate is 4800 bits per second and the payload payload PL-2 is 288 bits, i.e. 36 bytes. At the Node N in the base station, the payload PL-1, PL-2 of the call period received from the radio path is separated from other information transmitted on the radio5 path and a packet P1, P2 is made of the payload as explained in connection with Fig. 4. Correspondingly, the payload PL-1, PL-2 is taken from the packets received from the transmission network at the Node N, other necessary information is added to it on the radio path and the call period is forwarded to the mobile station in the correct time slot. Figure 3 does not show the exact detailed structure of the call period or the actual location of the data, but the representation only describes the data content of the call period so that the basic idea of ​​the invention can be clearly expressed. The hatched section in Figure 3 illustrates the payload moving over the radio interface and the black section the other payload moving on the same channel on the radio path, such as channel coding. In reality, in a call period sent over the air interface, the lined portion and the specified portion of the black portion are mixed together.

In addition to the packets shown for data transmission in Figure 3, the TETRA standard defines highly protected data with a data transfer rate at the air interface of 2400 bits per second, whereby the actual payload of the call period is 144 bits, i.e. 18 bytes. In addition to data transmission, the TETRA standard has two possibilities for voice transmission. Typically, speech is transmitted by transmitting it as channel coded speech, with a payload of 274 bits, or 34.25 bytes, per call period. It is also possible to transmit speech without channel coding, in which case the payload of one call period is 432 bits or 54 ta25 bytes. Thus, unencoded speech and unprotected data use exactly the same payload.

The data rates shown above are net data rates and the corresponding payloads are net payloads. This means that, for example, a data compression program could succeed in 864 bits of data in the pak30 bark to 432 bits, with a gross payload of 864 bits (and a gross data rate of 14,400 bits per second) but a payload of 432 bits at a net data rate of 200 bits per second. The data transfer rate can also be increased by reserving several channels for one connection. However, the net data rate refers to the rate used on one channel and the payload corresponding to the net data rate means the amount of data transmitted on one channel in one frame.

Figure 4 shows the structure of a communication packet according to the invention in a first preferred embodiment using the TET5 RA standard. The packet consists of a payload PL, the length of which in the first preferred embodiment is determined by the data rate. Its length is thus either 432, 288, 274 or 144 bits (i.e. the length in bytes 54, 36, 34.25 or 18). The invention is in no way limited to the use of these dimensions and, for example, a 432-bit payload could be divided into two parts and transmitted

288 and in 144-bit packets, or 274-bit payload, respectively, could be transmitted in a 288-bit packet. It is essential in the invention that the data is transmitted in packets of constant length, of which there are at least two constant lengths. By defining a separate transmission packet size for each data rate, i.e. the payload of the radio path, the advantage is achieved that the processing of the packets15 is facilitated and the operation of the network can be further optimized.

Referring to Figure 4, a header H is added to the payload, the length of which in the first preferred embodiment is a constant 4 bytes, i.e. 32 bits. The structure of the header is described in more detail in connection with Figure 5. Each payload length could also have its own header length, for example, a 24-bit header could be added to a 432-bit payload, a 32-bit header to a 288-bit payload, a 46-bit payload to a 274-bit payload, and a 40-bit header to an 18-byte payload. The advantage of using a single standard header is that it is easier to process packets at network nodes. The advantage of using your own header lengths is that you can adjust the packet sizes so that the optimization of network usage is diversified and the number of packet sizes to be transferred is reduced.

In the packet shown in Fig. 4, a terminal C is added to detect packet communication errors. In the first preferred embodiment, the terminal C is not added to packets with a payload of 432 bits. After all, this payload is shifted unprotected / unencoded, i.e. without error checking over the air interface, which is much more error prone. When using this data transfer rate, the correctness of the data transfer is not very important. The advantage of an endless packet is that the total length ly35 of the packet to be sent is slightly reduced, the packet is made to leave a little faster and the packet can be received faster because there is no need to check the packet for possible transmission errors. In the first preferred embodiment, the terminal is added to those packets with a payload of either 288, 274 or 144 bits. In the first preferred embodiment, the terminal C is the same in all three, for example a 16-bit cyclic check CRC (Cyclic

Redundance Check). Thus, in the first preferred embodiment, call packets P with a size of either 464, 336, 322 or 192 bits enter, move and leave the actual transmission network. In other embodiments of the invention, the terminal does not need to be used, or the terminal may be the same in all packets, or a separate terminal may be defined for each packet size. The terminal may also be enabling error correction, for example forward forward correction (FEC). It is essential for a packet according to the invention that its length is always one of a predetermined constant length and that it comprises at least a header of a predetermined length and a payload of a certain length.

The structure of the header of the first preferred embodiment is shown in Figure 5. The header always starts with the first bit, which identifies the start of the message packet. When no packets are transmitted in the transport channel, octets with a first bit of zero can be transmitted there. In other embodiments, multiple bits, different bit patterns, or a header with a zero bit may be used as the preamble of a message packet.

In the first preferred embodiment, the next bit is a bit

C / S, which indicates the type of packet, i.e. whether it is a control packet containing signaling information or a call or data packet.

The next 14 bits A tell the packet address. It can be the destination25 node address or a logical address depending on whether it is a control packet or an information packet. In the first preferred embodiment, some of the addresses are reserved as general addresses, by means of which, for example, changes in routing tables can be distributed to network nodes.

The TETRA standard defines many different voice and data services30 with different priorities. Priority determines how packets are handled in congestion situations. In the first preferred embodiment, the packet header multiplies the priority of the packet by the four bits PR following the address A.

With the next two bits in the header, the N packets are numbered sequentially to detect the missing packet. Bits N can also be used to tell how many slots, i.e. channels, a packet is intended for. Bit11 en N can also be used to transmit other information that network nodes may need to maximize network capacity utilization.

Bits N are followed by a two-bit bit sequence L, which is a packet length code. In the first preferred embodiment, the length code 00 means that the total length of the packet is 464 bits, the code 01 the total length 336 bits, the code 10 the total length 322 bits and the code 11 the total length 192 bits. This makes it possible to indicate the length of the packet with two bits. If the length of the packet were expressed without the length code in actual numbers, for example, the length 464 would have to be multiplied by a sequence of nine bits instead of the two-bit sequence according to the invention. Even if the lengths are expressed in bytes, a longer length sequence is required: 464 bits are 58 bytes and a sequence of at least six bits is required to express the number 58.

In the first preferred embodiment, the length code sequence L is followed by the last part of the header, an eight-bit cyclic checksum for detecting CRC header transmission errors. The advantage of using a cyclic checksum is that header transmission errors are detected and the packet going to the wrong address is not forwarded.

If a header is used in the embodiment of the invention to distinguish between data packets and control packets, for example, length code 00 can mean a data packet of 464 bits or a control packet of 2048 bits (256 bytes). The header bit C / S allows the network node to know which size is in question. Of course, an option in which the length code 00 always means a packet of the same size, for example 464 bits, is also possible. In embodiments using the header bit C / S, the control packets may also be of variable length, in which case, for example, the first octet of the content, i.e. the payload, indicates the length of the control packet, preferably in bytes. It can also use the space reserved for the header checksum and length code in applications where control packets are received in full before being forwarded. In this case, the control packet ends with, for example, a double-byte checksum, and before forwarding, it is checked whether the entire control packet has been transferred correctly.

In some embodiments of the invention, possibly "leftover" data packet length codes may indicate that the packet is transmitting constant length signaling information. Length codes are left over when the number of payloads35 of the data packets differs from the powers of the two, as could be the case in a TETRA system for data transmission only, where the payload108902 en sizes are three. There, the remaining length code 10 could indicate, for example, a packet containing 64-bit signaling information, regardless of whether bit C / S is used or what the value of bit C / S is.

The structure of the header shown in Figure 5 is indicative and not all parts of the header shown in Figure 5 are required in all 5 embodiments. In the header according to the invention, it is sufficient that at a certain point in the header of the data packet a bit sequence L is found, which contains the packet length code. The length of the bit sequence L in bits preferably depends on how many payload lengths are defined. All of the above figures are indicative and are intended to illustrate the invention without limiting it thereto.

Figure 6 shows a block diagram of a network node according to the invention in a first preferred embodiment. In the example, the network node is a selected TETRA exchange DXT, which has three input channels and from which two output channels originate, as shown in Figure 6. The only limitation on the number of output and input channels of the network node 15 according to the invention is that the network node has at least one input channel. The network node DXT comprises central terminals for sending and receiving ET packets.

The network node also comprises a call control unit CCU, which takes care of call control and resource management. The call control unit

The CCU controls the sending and receiving of packets through the central terminals ET. The call control unit CCU recognizes the information in the header of the received packet. In the first preferred embodiment, the call control unit CCU detects on the basis of the C / S bit whether it is receiving a data or control packet. The call control unit determines the length of the packet to be received25 from the length code in the packet header and knows from it because the whole packet has been received. To convert the length code to the correct length, the call control unit CCU may, for example, maintain a conversion table T1 with a code column L 'and a length column 3' indicating the actual length 3 corresponding to the code L. Table T1 may also contain other columns. For example, if packets with full byte lengths are to be transmitted in the TETRA transmission network, the conversion table T1 may contain the column "length in bytes" in addition to the columns described above. In this case, for example, packets containing the same number of bytes can be used for channel-coded speech and protected data, the rest of the channel-coded speech is only padding bits. The information in the table allows the call control unit CCU to know when the padding bits start and when the entire packet has been received. Other ways can be used to convert the length code L to the actual length 3.

In the first preferred embodiment, the central terminals ET also receive and send packets to other networks. Based on the input channel, the call control unit CCU knows whether it receives a packet from which the transmission network packet according to Fig. 4 has to be modified or whether it receives a ready-made packet from the transmission network. If the packet is received from the transmission network, the call control unit CCU checks whether the header of the data packet received from the transmission network has been transferred correctly by calculating the header checksum and verifying it with the checksum in the header. If the numbers are the same, the header has shifted correctly and the call control unit CCU starts examining the address in the packet header to send the packet. The call control unit CCU routes the packet using conventional routing methods, for example routing tables. If the packet received from the transmission network is sent back to the transmission network, the call control unit CCU may instruct the terminal ET to start sending the packet to the correct address, even if the packet has not been completely received. If the previous packet is currently being sent to the same output channel, the packet will have to wait a moment. However, it is forwarded substantially immediately, i.e., its transmission begins as soon as the previous packet is completely transmitted. This has the advantage of making the packet transmission time as short as possible.

If, in the first preferred embodiment, the call control unit CCU detects based on the address of a packet received from the transmission network that the packet is no longer being transmitted to another node in the transmission network, it unpacks the packet and For example, if the received packet contained a checksum for error checking, the call control unit checks whether the packet has been transferred correctly. Since in the first preferred embodiment the packet transmission errors are not checked in the transmission network, but are checked only when leaving the transmission network, in the first preferred embodiment the interruption of speech caused by the missing packet is prevented. In addition, the transfer is accelerated when the check is not performed on every node. The treatment of an incorrectly transferred packet is determined by the application used:

it can be sent incorrectly, deleted, replaced with a refill packet, or retransmitted.

O 8 90 2

If the header of the packet received from the transmission network has been transmitted in error, the call control unit CCU does not forward the packet or perform packet conversions.

If the received packet does not come from the transmission network, but is a modified packet, the call control unit CCU constructs a packet to be transmitted in the transmission network as described in Fig. 4, separating the payload, adding the header described in Fig. 5 with the correct information and possibly a checksum.

In the first preferred embodiment, the data packets are separated from the control packets on the basis of the C / S bit of the header shown in connection with Figure 5. When the network node DXT builds a transport packet, it places the bit C / S in either the data packet or control packet position based on which data it is. When receiving a transmission network packet, the network node knows, on the basis of the bit C / S, which packet it is. In a first preferred embodiment, the network node is adapted to receive and transmit variable length signaling packets having a header structure different from the header of the data packet. For example, in the exchange DXT, the call control unit CCU identifies the control packet and determines its length from the first byte of the payload.

In other embodiments of the invention, the network nodes may not have all of the above elements or, alternatively, their functions may be more versatile. It is sufficient for the system according to the invention to comprise at least two network nodes, one of which is able to at least receive the packet and identify the actual length of the chain from the packet length code and the other part at least add a header containing the length code to the packet and forward the packet.

It is to be understood that the foregoing description and the accompanying figures are intended only to illustrate the present invention. Various variations and modifications of the invention will be apparent to those skilled in the art without departing from the scope and spirit of the invention as set forth in the appended claims.