FIELD OF THE INVENTION

The invention relates to a method for packet communication in a radio access network.

The invention also relates to an arrangement for packet communication in a radio access network comprising a base station controller and a base station transceiver connected by a physical packet communication medium. The invention further relates to a computer program product comprising a computer-readable medium with program code that when read and executed by a computer causes the computer to execute such method.

BACKGROUND OF THE INVENTION As in recent times the widespread use of mobile devices becomes increasingly important in the daily life the associated infrastructure to support data and voice traffic has to adapt to the capacity requirements that are a consequence of these developments. At the same time the demand to stay competitive among the suppliers of such infrastructure requires it to rely as much as possible on standardized solutions available on the market e.g. for computer networks. In this context the GSM EDGE radio access network has been developed and evolved and the standards of the GERAN will be maintained by the 3GPP (3rd Generation Partnership Project) which is a key part of GSM and also of combined UMTS/GSM networks. The evolution in the area of the core network has led to a transition from the transmission between the base station controller and the base transceiver station of the core network taking place over time division multiplexing over the Abis interface to transmitting the corresponding data over a packet network installed between the base transceiver station and the base station controller. The general aspects of the associated infrastructure topology are well-known in the art and are standardized by the 3rd Generation Partnership Project and evolve in an ongoing process as documented on the homepage on the 3rd Generation Partnership Project under for instance http://www.3gpp.org/gsm-edge-radio-access-network. Due to increased bandwidth requirements as explained above there exists a need to improve the data transmission capacity between the base station controller and the base transceiver station in order to be able to support a higher number of mobile subscribers with a given infrastructure while managing data and voice traffic between the base station controller and the base transceiver station over a packet network. Other than the standards of GSM and 3GPP no particular prior art in this area is known.

SUMMARY OF THE INVENTION

It is an object of the present invention to improve the packet communication between a base station controller and base transceiver station in a radio access network.

This problem is solved by a method for packet communication according to claim 1 , an arrangement for packet communication according to claim 13, and a computer program product according to claim 15. Further advantageous embodiments of the invention are given in the dependent claims.

Said method achieves transmission of a plurality of communication channels in frames (600) between a base station controller (140) and a base transceiver station (160) by carrying out following steps: in a first grouping step, grouping a first number of frames associated to at least a first communication channel DSO ₀ in a first fixed time period (205) in a first group for transmission in a first packet (325); in a first transmission step, transmitting the first packet and maintaining the first packet at a transmitter and at a receiver; performing a second grouping step for the second fixed time period for grouping the at least first communication channel DSO ₀ in a second group; comparing the first and the second group and in case they are identical transmitting first identity information in the second packet, and at the receiver upon reception of the first identity information replacing the second group by the first group in the second fixed time period.

Said arrangement comprises thereto a base station controller (140) and a base station transceiver (160) connected by a physical packet communication medium (155), the base station controller and the base transceiver station. It further comprises a grouping entity for grouping in a first grouping step a first number of frames associated to a first communication channel (DSOO) in a first fixed time period (205) in a first group for transmission in a first packet (325); The grouping entity is further configured for performing a second grouping step for the second fixed time period for grouping in a second group. The arrangement additionally comprises a transmitter for transmitting in a first transmission step the first packet and for maintaining the first packet at the transmitter, and a receiver for receiving said first packet and for maintaining the first packet at the receiver. Furthermore, a comparator is present for comparing the first and the second group and in case they are identical transmitting first identity information in the second packet. A replacing entity is present for replacing at the receiver upon reception of the first identity information the second group by the first group in the second fixed time period. Advantageously the method according to the present invention allows to group a plurality of frames that are for instance transmitted during a packet transmission period. The associated group is stored at the transmitter and at the receiver. In this manner, the grouped frames of one time period can be compared with the group frames of a next fixed time period. A judgement upon the identity of the subsequent frame groups is based on this comparison.

Consequently, instead of transmitting the information of the first group again as part of the second group in the next frame, merely an identity information is transmitted to the receiver. This identity information is used at the receiver to duplicate the first group. The receiver will then further use the first group frames stored there instead of the second group of frames.

In this manner it is possible to avoid the transmission of redundant information between the base station controller and the base transceiver station. Therewith, the packet communication network may be used more efficiently. Preferably, the transmitter forms part of the basestation controller and the receiver forms part of the basestation transceiver. In this case, the further use of the receiver, for instance a transmission to a mobile handset. It is however not excluded that alternatively or additionally, the transmitter is part of the basestation transceiver and the receiver is part of the basestation controller. Expediently according to a further embodiment of the method according to the present invention, frames of a plurality of communication channels can be grouped in a particular time interval. The grouped frames are then transmitted over the packet network in order to exploit the redundancy of a plurality of communication channels in one transmission step. In this manner a plurality of communication channels can be transmitted in one communication packet while at the same time each individual channel can be marked in the packet concerning his identity information.

Beneficially according to a further embodiment of the method according to the present invention a communication channel of a TDM frame is transmitted and the fixed time period is accordingly selected in order to minimize the conversion effort for handling GSM TDM frames.

Expediently according to a further embodiment of the method of the present invention confirmation information is transmitted to the transmitter from the receiver upon reception of a communication packet. This guarantees that no information is lost in the course of the transmission of communication channels according to the present invention.

Beneficially according to a further development of an embodiment of the method according to the present invention the confirmation information comprises a packet sequence number thus ensuring, that an appropriate packet may be allocated and if one packet in the sequence is lost, the proper packet may be retransmitted from the transmitter to the receiver.

Expediently according to a further development of an embodiment of the method according to the present invention a communication packet from the base transceiver station to the base station controller is used to transmit the packet sequence number which allows a further reduction in the traffic associated to performing the method of the present invention, as normal communication packets for instance transmitting uplink traffic can be used to convey the information associated to performing the method according to the present invention.

Expediently according to a further development of the method according to the present invention the identity information is transmitted in form of a compression header indicating a list of not transmitted communication channels as this represents an efficient method to communicate which information can be duplicated on the receiver side from the packets stored there after receiving them during previous transmissions.

Advantageously according to a further development of an embodiment of the method according to the present invention a plurality of communication channels are acquired during a fixed time period as a chunk, as this allows it to adapt the method according to the present invention to the size of the GSM TDM frame precisely representing the number of transmitted communication channels and the corresponding 20 ms transmission period. Expediently according to a further development of the method according to the present invention the transmission occurs over Ethernet, which is a widely accepted and used packet communication network that is available on the market at competitive prices and at the same time operates in a highly reliable manner. Advantageously according to a further development of the method of the present invention the base station controller and the base station transceiver are connected over an Abis interface and thus the method according to the present invention is applicable to the common standard environment of the GSM EDGE radio access network. Expediently according to a further development of the method of the present invention once no confirmation information is received the subsequent packet is transmitted containing the frames of all the communication channels like the first packet is.

Expediently according to a further development of an embodiment of the method according to the present invention once no confirmation information is received it is evaluated at the transmitter side, if the information to be transmitted in the subsequent packet is redundant with information that has been previously transmitted with another packet and confirmed, and if this is the case such information is not transmitted and only the identity information is transmitted instead with the subsequent packet. In this manner it is possible to further optimize the use of the communication path between the base station controller and the base transceiver station without sacrificing any accuracy by more efficiently exploiting the knowledge about the information that has been transmitted and confirmed and is stored at present at the transmitter and at the receiver.

Advantageously the arrangement for packet communication according to the present invention comprises all the communication entities and resources that are required to perform the method according to the present invention and thus allows to perform the method with a minimum number of hardware resources.

Expediently according to a further embodiment of the arrangement of the present invention the physical packet communication medium is arranged as Ethernet and thus allows to use standardized packet communication networks in the arrangement of the present invention.

Beneficially the computer program product according to the present invention allows it to easily store and distribute a computer program code for executing the method of the present invention as process steps in various devices such as base station controllers and base transceiver stations of a radio access network.

BRIEF DESCRIPTION OF THE FIGURES

In the following the invention will be further described on behalf of examples and embodiments represented in drawings, wherein:

Fig. 1 shows a configuration of a radio access network,

Fig. 2 shows an example of a channel usage in a TDM link,

Fig. 3 shows an example of a TDM channel usage according to an embodiment of the present invention, Fig. 4 shows an example for extraction of columns from TDM frames,

Fig. 5 gives an example of packet confirmation according to an embodiment of the present invention,

Fig. 6 shows an example of an RTP packet format, Fig. 7 gives an example of a packet header, Fig. 8 gives an example of a sequence of packet confirmations,

Fig. 9 shows an example for compression cancellation, and Fig. 10 gives an example for an optimization of compression cancellation. DETAILED DESCRIPTION AND EMBODIMENTS

The present invention will be described with respect to particular embodiments and with reference to certain drawings but the invention is not limited thereto but only by the claims. The drawings described are only schematic and are non- limiting. In the drawings, the size of some of the elements may be exaggerated and not drawn on scale for illustrative purposes. The dimensions and the relative dimensions do not correspond to actual reductions to practice of the invention.

It is observed for clarity that the terms first, second, third and the like in the description and in the claims, are used for distinguishing between similar elements and not necessarily for describing a sequence, either temporally, spatially, in ranking or in any other manner. It is to be understood that the terms so used are interchangeable under appropriate circumstances and that the embodiments of the invention described herein are capable of operation in other sequences than described or illustrated herein.

Fig. 1 shows an example of a radio access network. This network may be comprised of a public switched network 105, which is connected to a mobile switching centre 115. This mobile switching centre 115 is further connected to a telecommunication control unit 125, which is again connected to a base station controller 140. As indicated by the dotted line, the base station controller 140 was previously connected to the base transceiver station 155 via a TDM - time division multiplexing - connection (150) that realized the Abis interface. The radio access network further comprises an "Agprs"-interface between the basestation controller 140 and an internet or intranet 110. This "Agprs"-interface is embodied with a serving GPRS support node 120 - on the side of the internet 110 - and with a packet control unit support node 130 - linked to the base station controller 140. The GPRS support node 120 and the packet control unit support node 130 are coupled to each other. The basestation controller 140 is further coupled to a base transceiver station 140 over a packet network 145 that realizes the Abis interface.

The configuration as shown in Fig. 1 allows a smooth transition between the time division multiplexing communication between the base station controller and the base transceiver station 150 and 155 on the one hand, and a packet communication link 145 between the base station controller 140 and the base transceiver station 160, on the other hand. As will be explained further below, the packet communication link 145 can be used more efficiently.

Fig. 2 gives an example of a prior art approach of transmitting communication channels in packets. Herein, the packet communication is carried out in accordance with a E1 TDM link. Reference sign 205 indicates the 256 communication channels that may be transmitted per E1 TDM link. Reference sign 210 indicates 20 ms of speech whereas reference numeral 215 marks the DTX header for discontinuous transmission. Reference sign 200 marks the channel usage example for TDM. Reference sign 220 indicates an LAPD message to transmit control information. Furthermore reference sign 270 indicates space that is unused in the packets during transmission. Such an unused space may be for instance space that is unused due to half rate communication. As can be seen in the example of Fig. 2 with the TDM over packet approach all TDM content of the 256 8 Kbit/s channels for E1 is transmitted without any optimization. The inventors of the present invention have however recognized, that much of the transmitted information is not helpful, because many bits in TDM frames are unused. For the purpose of optimization it would be possible and feasible to analyze each respective communication channel for its potential of compression. This would be however a very complex and computationally intensive task, as an Abis frame contains multiple non- standardized formats. Such compression analysis therefore requires a complex software solution that would be highly proprietary. Use of a proprietary software solution is not desirable. Reference signs 225, 230, 235, 240, 245, 250, 255 and 260 point to the respective packet contents of the packets that are presently assigned to the TDM channels.

Fig. 3 gives an example of a TDM channel usage 300 according to an embodiment of the present invention. Again reference sign 305 indicates the 256 channels per E1 TDM link, 315 marks the DTX header and 310 marks 20 ms of speech. Reference sign 320 is associated to a control LAPD message and reference signs 325 to 360 in steps of five indicate respective contents of packets 1 to 8. Reference sign 312 marks the greyscale for optimized content. Reference sign 316 marks the greyscale for non-optimized content. Reference sign 318 stands for content that is removed from the packet due to redundancy. This TDM channel usage 300 allows a better optimization. For instance, when half rate communication takes place, most of the contents of packets 5 to 8 is unused as indicated by reference signs 376, 378 and 374. Consequently this content may be removed from the packet.

In one embodiment, the removal is carried out by exploitation of the features of GSM. Generally, packets are transmitted within a 20 ms period nature of a GSM TDM frame. All of the transmitted packets are memorized, both at the transmitter and at the receiver. Upon transmission, the transmitter compares a new GSM TDM frame to be sent with a previous one sent 20 ms before on a channel by channel basis. As a result of this comparison, the transmitter may reduce the channel size: identical content channels are not sent. Instead, a specific compression header is added which contains the list of non-sent channels. These non-sent channels then can be reproduced from the stored content at the receiver and the information in the compression header on the receiver side and be duplicated. Bandwidth on the packet communication link can be saved in this manner, as not the whole channel content has to be transmitted and only an identity information needs to be included in the header. According to one embodiment of the present invention, a static compression is used. This static compression represents a first level of compression that directly results from the definition of dialog protocols between the base station controller and a base transceiver station. Indeed for each PCM link serving one IP module of the base transceiver station and one IP gateway of the base station controller a first and a second channel are defined on the packet switched network. The GSM signal which conveys the LAPD message on a UDP/IP Sec connection, and the GSM traffic channels which convey TRAU frames on an RTP _N /UDP/IP Sec connection. The content of all DSOs of the PCM link which are not configured - neither as traffic nor as LAPD DSOs - is not transmitted on the packet switched network which is called a static compression.

The present invention in this embodiment proposes dynamic compression in the form of differential compression for traffic channels. According to the present invention the differential compression for traffic channels serves to save bandwidth on the packet network by not transporting most of the useless information which is carried by some TRAU frames as there is an idle bit pattern, GPRS idle frames, and a part of the DTX frames. The advantage of differential compression is that it works without any knowledge of the kind of traffic in terms of voice, circuit-switched data or GPRS that is transmitted in the traffic channels. The configuration of channels - full rate or half rate - may also be unknown. Differential compression exploits the fact that when a traffic channel is in a silent situation meaning in an active circuit-switch channel or an idle GPRS channel the same information will be repeated every 20 ms. In some other cases like in DTX (discontinuous transmission) consecutive frames are partially identical. The differential compression consists in comparing the data being transmitted with the data of the same channel transmitted 20 ms before. Advantageously the compression function operates on chunks of traffic. Those chunks are made of the content of all traffic DSOs of a given PCM link for a certain time interval which is preferably equal to the period of packet transmission T _PT .

Preferably to make a compression operate properly this time interval is selected to be a divider of 20 ms. A chunk of traffic is preferably divided in columns. A column is made of the content of a particular 8-kbit/s channel inside traffic chunk. Each column is then compared with the column corresponding to the same channel in the chunk transmitted 20 ms before. A chunk used for such a comparison may preferably called reference chunk. A column can be compressed or not according to the following rule: 1. If the content of the column is identical to the content of the same column in the reference chunk, then the column is compressed which means that the data of the column are not included in the transmitted packet.

2. If there is a difference between a column and the one of the reference chunk, the column is not compressed meaning that the data are included in the transmitted packet.

For some IP networks with higher transfer time a differential compression method can be applied by comparing each chunk with a reference chunk transmitted D ^* 20 ms before (40 ms or 60 ms ...). A value of D strictly greater than 1 does not include any loss of compression capability for a channel permanently in "silent" state. A compression may be performed for LAPD channels transmitting control information as well as TRAU frames transmitting voice information. A static compression is for instance communicated between the base station controller IP gateway and the base transceiver station IP module in form of the DSO mapping of the IPG-IPM link which is provided in a TCP message.

Fig. 4 gives an example of an extraction of columns as an example for groups from TDM frames. According to an embodiment of the present invention the same compression algorithm may be implemented in the transmitting entity of the base station controller IP gateway for downlink and the base transceiver station IP module for uplink traffic.

Fig. 4 for instance shows an example of a chunk of traffic channels where a plurality of TDM frames here N TDM frames 430 are acquired over a fixed time period. The traffic channels DSO ₀ 410, DSOi 415, DSO _j 420 and DSOm-1 425 are marked. Traffic channel DSOo is represented in a column 450 and traffic channel DS0 _m- i is represented in a column 470. For instance, in a first step of compression a chunk of traffic made of N TDM frames during a period of time equal to the packet transmission period is acquired. Each of these TDM frames contains m traffic DSOs. The number of traffic DSOs and their position are defined in the DSO mapping provided in the IP service channel. While a TDM frame is delivered every 125 μs, 8 frames are delivered per ms. As a consequence of the packet transmission period being an integer number of ms, N is a multiple of 8. Furthermore the data of a chunk are then organized in M=8 ^* M columns of N=n/8 bytes as illustrated in Fig. 4. Here column no. j indicated by 420 (O≤j≤M-1 ) is made of bits C(ij) (O≤j≤N-1 ) of a single 8-kbit/s channel. The content of each column C(J) is compared with the same column in the chunk transmitted D ^* 20 ms before, in order to determine if the column is to be transmitted or not. Variable D denotes the compression memory depth which is a parameter for the number of frames that are kept in memory. Columns which are identical are e.g. removed from the transmitted packet which will then contain the set of columns which are not identical and the compression header. This header is for instance made of M bits wherein 1 bit corresponds to each column. If this bit is for instance set to 0 the column may be identical to the one transmitted D ^* 20 ms before and if it is 1 the column is presently transmitted. Furthermore it is preferable to store the uncompressed chunk in memory as well as in the transmitter and in the receiver in order to be used as a reference chunk.

Fig. 5 shows an example of an acknowledgment mechanism according to an embodiment of the present invention. Reference sign 505 marks the IP gateway of the base station controller and reference sign 510 marks the IP module of the base transceiver station. Whenever the IP gateway or the IP module for instance receives an RTP packet with a sequence number SN it tags the next transmitted packet on the same link with a piggy-backed sequence number in the S.N. ACK field in the RTP _N header. The piggy-backed sequence number for instance is an 8-bit field equal to PB(SN)=SN Modulo 256. For instance, packet transmissions 520, 530, 540 and 550 are marked with their normal sequence numbers for uplink and downlink SN1 _D L, PB(SNOUL)] and so on marking the normal packet transmission flow whereby each of the 16-bit sequence number SN _DL and SN _UL are independent from each other and are managed respectively by the IP module of the base transceiver station for uplink numbers and by the IP gateway of the base station controller for downlink sequence number. The piggy-backed sequence number is corresponding to the last sequence number SN (or the one with the higher sequence number in case of a disordered arrival) which has been received between the previous transmitted packet and the next one to transmit. If no packet is received during this interval the piggy-backed sequence number provided will remain the same as the previous one provided and a control byte will be set to 0 informing the receptor of the invalidity of the piggy-backed sequence number. The acknowledgement mechanism is based on the fact that when a packet with a sequence number SN has been transmitted the transmitter expects a packet in the opposite direction with the corresponding piggy-backed sequence number to be received within the maximum round-trip-time of the packet switched network. If the piggy-backed sequence number has not been received at that time, the packet is considered to be lost and present chunks will not be used as reference for the compression. On the other hand, if many packets have been transmitted from the receiver the IP transmitter expects a packet in the opposite direction to be able to supply a new piggy-backed sequence number in the new packet to transmit. As explained above, the piggy-backed sequence number transmitted is the last or higher sequence number which has been received and if no packet is received, the IP transmitter will input the previous piggy-backed transmitted sequence number and set to 0 an acknowledge bit of a control byte informing of the invalidity of the piggy-backed sequence number.

Fig. 6 shows an example of a real-time packet format 600 which is transmitted over the packet network between base station controller and base transceiver station. Here it comprises an Ethernet header 605 and a compressed payload 625 which is comprised of columns 613, 618 and 628 and at 623 of a list of columns that are not transmitted because the content of these columns is identical to the ones that have been transmitted with the previous packet. Upon reception of such a real-time packet the receiver will arrange the packets in the corresponding sequence and take the ones from the reference chunk to fill in the empty space of the payload to compose a complete set of communication channels.

Fig. 7 shows an example of an RTP _N header 700, 715 is the location of the piggy-backed confirmation sequence number whereas 710 is the location of the normal packet sequence number. 720 represents the DSO bit map.

Fig. 8 explains an example of potential packet network impairments and a corresponding behaviour of the receiver regarding a validity bit of the sequence number. Here the emitter 805 communicates packets with a receiver 810. Here packet SN#1 and SN#2 arrives in the same window. The next piggy- backed sequence number supplied by the receiver is at 823 the last higher sequence number PB(SN#2). A piggy-backed sequence number for packet SN#1 is never sent so at 83 the emitter 805 has received a piggy-backed sequence number 2. Consequently here the emitter 805 acknowledges the reception of piggy-backed sequence number #2 at 833. For packet numbers 3 and 4 no new packet is received so the corresponding piggy-backed sequence number remains unchanged while the validity bit is set to 0 to inform that the piggy-backed sequence number is invalid. With 85 a packet having the sequence number 5 is sent and received at 838 which leads to the transmission of the piggy-backed sequence number 5 from the receiver 810 to the transmitter 805 at 886. Here similar to the case above packet number 3, 4 and 5 arrive at the same window even if packet number 5 arrives before packet number 4 the next piggy-backed sequence number applied by the receiver is the higher sequence number received so it is PB(SN#5) and piggy-backed sequence number for packet SN#3 and SN#4 respectively coming from a 84 and 83 are never sent.

Fig. 9 gives an example for a data flow to effect cancellation of compression. Such an evaluation is preferably performed in order to avoid complications as a consequence of packet losses. This is particularly important, in case whenever a reference chunk has not been acknowledged. In case a packet is about to be transmitted the IP transmitter checks preferably the acknowledgement state of the reference chunk 905 at an evaluation 910 before proceeding with compression, meaning omission of certain packet content. If the packet containing the reference chunk has been lost or is still waiting for acknowledgement compression of the columns is prohibited at 913 respectively 923 and an RTP packet which is uncompressed is sent at 928, meaning that all the frames are contained in the respective columns to be transmitted in the packet. On the other hand if the reference chunk has been acknowledged at 918 compression will be effected at 938 allowing a packet to be sent, which only contains the modified columns at 943 and the compression header indicating the identical columns to the previously transmitted chunk.

Fig. 10 gives an example for compression optimization according to a further embodiment of the present invention. For instance, instead of a prohibited compression of all columns in the case the compressing side did not receive an acknowledge of the reference chunk, there is an alternative way to optimize the compression procedure. This alternative method is based on the principle that the compressing side may not compress columns that were included in the non-acknowledged packet. According to that the packet still can be partially compressed. This is an important advantage in case the network load increases during busy hours. In this case the delay and jitter can lengthen with a risk of loss on compression packet acknowledges. Thus this optimization allows to avoid excessive packages with maximal size where the packet is not fully compressed, which are unfavourable for the network.

As shown in Fig.10, a packet exchange 1000 between an IP module 1020 of a base transceiver station and an IP gateway 1010 of a base station is shown. In a first step a packet 1030 is sent from the IP module to the IP gateway and at 1033 is acknowledged by a piggy-backed sequence number from the IP gateway 1010 to the IP module 1020. Greyscales 1023 indicate a column that may be compressed, 1018 a column that is not compressed and 1015 a column which is compressed. In the beginning all columns have the colour 1018 and thus none of them is compressed. In the second transmission step packet 10b is transmitted in 1038 to the IP gateway and acknowledged in 1043 to the IP module. As columns 2 and 4 were the same as in case of packet 10a those columns are compressed marked by colour 1015. Furthermore packet 10c contains compressed columns 1 and 2 where column 1 is compressed and column 2 which is identical to packet 10b remains compressed. In this case the column 4 has changed and thus is transmitted uncompressed marked by colour 1018. In this case however the IP module 1020 receives no acknowledgement packet and transmits packet 10b as a consequence of the missing acknowledgement packet. Here as columns 3 and 4 have not been acknowledged they have to remain uncompressed. On the other hand columns 1 and 2 are the same as in packet 10c and thus may be compressed as indicated by colour 1023 because they have been already acknowledged previously.

Further to this procedure there also exists the possibility to work on a frame nearly identical hypothesis. This means, once a packet is lost the respective receiver after passing of the round trip time may automatically generate another packet which is identical to the one that has been received before to fill in the gap. Each receiver may also take provisions to store enough packets to be able to allow a reordering of the packets in the proper sequence number in order to process them in a meaningful manner.

In short, the present invention relates to a method for packet communication in a radio access network. The method leads to an increase in the capacity of a packet switched network between a base station controller and a base transceiver station. This is obtained by gathering a number of TDM frames associated to communication channels to be transmitted in a TDM frame. This occurs in form of a chunk transmitting them, while at the same time storing them at the transmitter and at the receiver. Upon the next transmission, columns representative of a respective communication channel of a next transmission period are compared with a previous one. A determination is made if identical columns are contained. If so, these columns are omitted from transmission. Instead, a reference is transmitted in a compression header indicating the number and position of identical columns. This allows the receiver to generate a complete number of communication channels, e.g. by filling in the previously stored columns that have been not transmitted.

In this manner a higher number of mobile devices can be supported, as no idle patterns or silence patterns need to be transmitted over the Abis interface on the packet network. An associated acknowledgement scheme avoids problems associated to packet losses.

The invention further relates to an arrangement in which the method can be implemented, as well as a computer program for carrying out the method.