CN1256062A - Method for packet switched data transmission - Google Patents

Abstract

The invention relates to packet-switched transmission of calls and particularly to real-time packet-switched transmission of speech calls. The object of the invention is to provide a method that allows efficient use of the transmission capacity without impairing the quality of real-time calls: information can here be transmitted by a packet-switched method so that the transmission delays remain short and the transmission delay of one and the same call does not vary at the receiving end. The method of the invention comprises defining a repetition period (R) that is notably longer than the duration of the transmission of a single packet, defining the starting point (0) of the repetition period, receiving a packet (2), allocating a time slice (1') of the same duration as the packet from the repetition period for the connection that the packet is associated with, and transmitting the packet in the time slice allocated and when a new packet associated with the same connection is received, transmitting the packet in the time slice allocated for the connection. Another object of the invention is to provide a network node in a data transmission network.

Description

A packet switching data transmission method

The present invention relates to packet switching transmission of calls, in particular to real-time packet switching transmission of voice calls on circuits of telecommunication networks in mobile phone systems.

In most digital mobile telephone systems, the carriers of the radio path are divided among several users by methods such as TDMA (Time Division Multiplexing) or CDMA (Code Division Multiplexing). The commonality of these methods is that the call is divided into call packets of a certain length, which are transmitted in frames on the radio path. Calls are typically routed using a circuit-switched system on circuits in the network infrastructure. If digital speech coding such as ACELP (Algebraic Code Excited Linear Prediction) is used on the radio path, then the packets received from the radio path in packet form must be disassembled and, in turn, assembled when sending them on the radio path .

In circuit-switched networks, each call is assigned a different transmission channel, and even a different channel in each direction of transmission. In a circuit-switched system, channels remain assigned even when no information is being transmitted, making them less efficient. The efficiency of circuit-switched systems can be increased by reducing the transmission rate. However, low transmission rates increase transmission delays. In mobile telephony systems where information is transmitted in packets over a radio path, the relevant delay is that experienced by the last bit of the packet, since a packet can only be encoded when the entire packet is available. transmission. In this way, a packet may miss its allocated time slot, the transmission capacity is not effectively utilized, and the quality of the call will suffer.

Another problem in circuit-switched networks is that as the load increases, capacity will terminate abruptly, whereas in packet-switched networks, an increase in load will only result in a reduction in the transmission rate and a deterioration in the quality of the connection.

In a packet-switched network, the arrival rate of packets at the receiving end varies, for example, with the load of network nodes en route. If the transmission period of packets at the receiving end, ie, the arrival rate, varies too much, the quality of speech will suffer when it comes to voice calls. In speech transmission, it is important to keep the maximum delay short and constant throughout the speech item. In other words, the transmission delays of voice call packets must be as uniform as possible. Furthermore, variations in the transmission time of data call packets in a mobile telephone system are also undesirable, because according to the radio protocol, call packets must be sent at an equal rate to the radio path.

One way to overcome the above-mentioned problem, ie, the variation of transmission times in packet-switched networks, is by transferring packets in a manner that mimics circuit switching, ie by assigning them separate channels. The problem with this scheme is the same as with the circuit-switched scheme, namely that capacity is allocated even though it may not be needed, and that capacity will be terminated abruptly as load increases.

Another way to equalize the transmission delay of packets in a packet-switched network is to use a transmission channel with a very high rate, so that the transmission time of packets is short compared to the interval between packets. The problem with high-speed transmission networks is that the network costs are high, and if the different channels of the transmission network have different rates, the voice packet must be fully received before the voice packet can be sent to the faster channel. This results in additional delay compared to the case where the entire transport network is operating at the same rate, so that the beginning of the packet arrives before the end of the packet is sent.

Another alternative is to buffer a sufficient number of packets at the receiver so that any variation in transmission delay can be balanced out. The problem is that the real-time transmission of speech does not allow large buffers: in common safety and security calls, for example, the delay must be less than 400 microseconds. In the Internet, voice delivery software uses larger buffers, because a few seconds delay in voice transmission does not matter much there. Because in a system that transmits voice in real time, the voice must be connected as soon as possible, so the buffer must be very short, so the buffer cannot be relied on to equalize the transmission delay.

The object of the present invention is to provide a method and a network node which enable a packet-switched transmission network to transmit real-time voice. This object is realized by the packet switching data transmission method of the present invention, and this method is characterized in that, determine the repetition cycle that is far longer than the required time of sending a single packet, and allocate time slices of different lengths for the connection from the repetition cycle; determine the repetition cycle Start point; maintain the information of the current point in the repetition cycle; maintain the time slice information allocated from the repetition cycle, and the connection information associated with it; receive the packet; identify the connection and the destination address associated with the packet according to the identification data of the packet ; Check whether a time slice is allocated for the connection of the packet from the repeat cycle; if yes, then when the start point of the allocated time slice is reached in the repeat cycle, start sending the packet; if not, determine the packet sending duration; in the repeat cycle Search for an idle time period at least as long as the sending time; from the idle time period in the repetition cycle, allocate a time slice at least as long as the packet sending time for the connection of the packet; when the assigned time slice is reached in the repeating cycle At the start point of , start sending packets.

A connection here means a data transmission, or a call, a connection, or a connection for sending signaling information between network elements. In systems where calls are handled as individual speech items, a connection means a single speech item.

The invention also relates to a network node that can be used for the method of the invention. According to the invention the network node of the packet-switched transmission network comprises sending means for sending packets to the destination address, at least one incoming channel leading to the network node and at least one outgoing channel originating from the network node, characterized in that the node also including setting means for setting the duration of the repeating period to be much longer than the transmission time of a single packet; at least one timer for indicating the current point in the repeating period to send packets at regular intervals, the timer being cleared at intervals of one repeating period zero; identifying means for identifying the connection associated with the received packet and detecting whether the network node has received the packet associated with the connection before said packet, or whether the packet is the first packet associated with the connection A group; assigning means for searching in the repeating period at least as long as the packet transmission duration of the free time slot, in response to receiving the first packet associated with the connection, from the free time slot for the packet. The associated connection allocates a time slice at least as long as the packet transmission duration, the start point of which is indicated by a timer determining the start moment of the transmission of the packet associated with the connection; and control means for The packets associated with the connection are sent to the destination address in the time slot allocated by the associated connection such that packets associated with the same connection are sent from the network node one packet at a time at intervals of one repeating period.

The basic idea of the invention is that the transmission delay of the packets is normalized by assigning each connection a packet-sized transmission time slot in the network nodes and forwarding the connection packets associated with a single connection at specific intervals. In this way, the invention has the advantage that the transmission capacity can be used effectively without impairing the quality of the call, since the information can be sent by packet switching method, so that the transmission delay is kept short and the transmission delay of the same call will not be lost at the receiving end. Variety. The invention is most advantageous in simplex (half-duplex) communications where data can be sent in both directions, but only in one direction at a time. Packet-switched packets are provided and resources are allocated only in the direction in which information is being carried, whereas circuit-switched transport channels must be allocated in both directions, but only one channel can be used at a time. Thus, with the invention, in the most favorable case, the transmission capacity of the circuit can be doubled.

Another advantage of the invention is that no separate channel needs to be allocated for signaling, signaling information is transmitted on the same channel as other data. This further improves efficiency.

Another advantage of the invention is that it can be applied to considerably different systems and since the invention does not place any restrictions on the size of the packets, no idle gaps need to be sent in packets other than information. Packets of length different from the packet lengths of other connections may be forwarded to each connection.

In a preferred embodiment of the present invention, if no packet needs to be sent to the connection for two consecutive times, the network node reclaims the allocated time slice. This has the advantage that time slots are not allocated ineffectively and the entire capacity is efficiently used. Another advantage is that the amount of signaling is reduced, since no separate call-end control information needs to be sent to the network node in order to reclaim allocated resources.

In another preferred embodiment of the invention, the control packets containing the signaling information are separated from the call packets and are sent between the call packets as soon as there is sufficient time. Call packets contain voice or data. This has the advantage that less time-critical control packets 'give way' to give priority to more time-critical call packets, thereby enabling a more efficient use of transmission capacity. This further ensures that the call packet transmission delays are short and equal.

Preferred embodiments of the claimed method and network node are set out in the dependent claims 2 to 6 and 8 to 11.

Below in conjunction with accompanying drawing, describe the present invention in detail by preferred embodiment, in accompanying drawing

Figure 1 is a block diagram of a transmission network according to the TETRA system;

Fig. 2 illustrates the relationship between the grouping of the wireless path and the grouping sent in the transmission network;

Figure 3 illustrates the allocation of time slices on incoming and outgoing channels of a network node according to a first preferred embodiment;

Fig. 4 is a block diagram of a network node;

The flowchart of Figure 5 illustrates the division of packets into control packets and connection packets, and

The flow diagram of Figure 6 illustrates the operation of a network node.

The invention is described below by way of an example of a packet-switched telecommunication network designed to act as a transport network for a digital wireless network following the TETRA standard (Tan European Trunked Wireless Network) defined by ETSI (European Telecommunications Standards Institute). ). TETRA defines standards such as interfaces to other networks, air interfaces and interfaces to another TETRA network. However, the internal structure of the transmission network is not defined in the TETRA standard, so the TETRA network is a good network example. In addition, the TETRA standard defines very time-critical calling and network services, because the network defined by the TETRA standard is also designed to serve as a private mobile wireless network for government agencies. However, the present invention is not limited to wireless networks or other wireless networks. It is obvious to those skilled in the art that the present invention can be applied to networks based on wireless data transmission and other data transmission systems in fixed networks.

Fig. 1 shows an exemplary structure of a TETRA transmission network. A mobile station MS (mobile subscriber) communicates with a base station BS via a radio path Air. This radio interface is defined in the TETRA standard. Each base station BS comprises a node N connected by circuit to a TETRA digital exchange DXT of the fixed transmission network. The TETRA exchange DXT is connected to other exchanges DCT of TETRA and digital central exchange DXTc through fixed circuits, and DXTc is an exchange connected to other exchanges DXT and/or other central exchanges DXTc to provide alternative paths for traffic. In this application, the interface to another TETRA network is located at the central exchange DXTc, but it could also be located at other exchanges DXT. In this application, the standard-defined external interfaces to the public switched telephone network PSTN, the integrated services digital network ISDN, the private automatic exchange PABX and the packet data network PDN are located at one exchange DXT, but they can also be located at individual exchanges, for example. The TETRA transmission network also includes other interface and peripheral units not shown in this figure. They include, for example, network management systems and dispatching systems.

In a TETRA system or similar, calls are handled as individual speech items, but in this application a call represents a speech item and a call group represents a speech item group.

In a first preferred embodiment, the node N of the base station BS is connected to the exchange DXT via a 64 kbit/s circuit. The switch DXT is connected to the rest of the network at a higher rate, eg 2Mbit/s. In the first preferred embodiment, the high-speed channel is preferably used as a parallel 64 kbit/s connection, so that 32 packets multiplexed with each other can be sent in parallel. The situation is the same as when packets arrive from different directions: the order of the packets is random. However, packets associated with the same connection are placed in time order because they are sent consecutively and travel the same distance at the same rate. It is also possible to use high-speed channels directly without dividing them into parallel channels. The transmission rates are given by way of example only to illustrate the invention.

The TETRA standard defines different voice and data services with different priorities. For each registered user, a priority is stored in the database. The priority and quality of services affect eg the allocation of resources. One voice service defined in the standard is a group call, which is established by, for example, dialing a group number from a location and stating that the call is a point-to-multipoint call. If a group call is being set up, the group identifier is only sent to base stations BS within the area of the group. The identification of the group to which the mobile user belongs is stored in the mobile station MS. The mobile station MS checks the received group identity and, if it is found in the memory of the mobile station MS, connects to the call: in this example it is assumed that the switch or node that needs to replicate the group call is capable of replicating a sufficient number of group identities. The group is called and forwarded in a single packet so that a single packet can be sent to each base station BS within the area of the group to be forwarded to the mobile users of the group. In this way, the group call packet can be processed in the same manner as the traditional call packet in the first preferred embodiment of the present invention.

Figure 2 illustrates the relationship between the packets of the radio path and the packets sent in the transport network. If the information is sent over the wireless path, frames are sent consecutively on the carrier. A frame contains a carrier-specific channel or connection, so the packet is associated with a call on that channel to a receiving mobile station. The base station then receives successive frames in which channels are allocated for each connection. In currently used methods based on dividing the carrier into channels, the frames are TDMA or CDMA frames. The transmission duration of a frame depends on the wireless system used. For example, in the TETRA system, it takes about 60 milliseconds to send a TDMA frame, and about 20 milliseconds to send a frame in the narrowband data transmission realized by the CDMA method. Fig. 2 shows the packets C1, C2, C3 and C4 that need to be transmitted on the carrier according to the TETRA standard in the first preferred embodiment. The wireless interface Air of the TETRA standard is based on a carrier wave with a bandwidth of 25 kHz for time-division multiplexing of 4 channels. For simplicity, the figure only shows one carrier in one direction. A base station can have multiple carriers, and there can be carriers in both uplink and downlink directions. Usually one channel is allocated for each connection, but the radio system of the TETRA standard allows multiple time slots to be allocated for one connection. But for the sake of simplicity, it is assumed that one channel is allocated for one connection. A group call is also a connection. Referring to Fig. 2, assuming that voice without channel coding is transmitted on channel 1, the actual information content of the packet, ie the payload PL1, is 432 bits. The speech transmitted on channel 3 is channel coded speech and the payload PL2 of the packet is 274 bits. In node N of the base station, the payload received from the radio path is separated from other information transmitted on the radio path, and identification data ID is added to the payload. The identification data ID1, ID2 indicate the connection to which the packet is associated, ie the destination of the packet. From the data, the node identifies the data transfer connection associated with the packet. This identifier indicates the destination address of the packet, ie where it needs to be sent. Identification data is represented by, for example, 56 bits. Correspondingly, the node N deletes the identification data from the packet payload received from the transmission network, and adds other necessary wireless path information. Therefore, the size of the paging packet PA sent to or received from the actual transport network is 330 bits or 488 bits. The transmission times of these packets on the 64 kbit/s transmission path are about 5.2 milliseconds and 7.6 milliseconds respectively. About 15 milliseconds are reserved for each channel on the wireless path. A network node of the transport network, ie the exchange DXT in the first preferred embodiment of the invention, receives the packet PA and forwards it, preferably according to the identification data ID.

Fig. 3 illustrates the node operation of the packet-switched data transmission network in the first preferred embodiment of the present invention. This node is eg an exchange DXT of the TETRA transport network. The node may also be any other network node that contains at least some intelligence such that it is able to send received packets back in the same direction. The node does not need to do any kind of processing on the packet; all it needs to do is forward the packet. These packets contain a payload and an identifier according to FIG. 2 . For simplicity, it is assumed that in Figure 3, the transmission rates of incoming channels 1, 2, 3, 4, 5 and 6 are the same as the transmission rates of outgoing channels A and B. This has the advantage that the sending of packets can be started immediately when the reception of packets begins. If the sending rates are different, then the sending cannot start until the entire packet has been received. In Fig. 3, the different data transmission connections are numbered in the order of the incoming channels, separating them from each other. Packets associated with the same connection have the same number. The numbered functions are purely illustrative. The numbered data transmission connections have been allocated a time slice 1 in a repetition period R having a predetermined length RT. According to FIG. 3, the repetition period must be longer than the transmission time of a single packet, so that there is sufficient time within the period to send packets of different connections. If the control packets containing network internal signaling information are separated from the call packets, then the length of the control packets need not be taken into account if the length of the repetition period is defined. Control packets are usually short, but larger control packets may need to be used, for example, when updating the network, so the determining factor for the maximum length of the control packet is that the transmission of the control packet must not exceed the duration of the repetition period. If the invention is applied to a wireless system, it is natural to define the length of the repetition period to be equal to the transmission duration of the frame on the wireless path, a packet in each frame requires a connection, because the packets associated with the same connection are always on the same channel to send. If the length of the repeating period is equal to the transmission duration of the frame on the radio path, the operation of the whole system is synchronized. If the repetition period is shorter than the sending time of the frame, the packets still cannot reach the receiving end sooner, but are buffered at the base station, where they wait to be connected to the wireless path. If the repetition period is long, some packets cannot arrive and be sent to the radio path in time, but miss their transmission slots, thereby impairing the quality of the connection. Therefore, in TETRA networks, the length of the repetition period is preferably 60 milliseconds. At the 64 kbit/s transmission rate used in this example, the length of the repeat period is approximately 11.5 times the time period required to send the shorter 330-bit paging packet, or 8 times the time required to send the longer 488-bit packet period.

In FIG. 3 , packet 2 arrives on ingress channel 6 associated with a connection to which time slice 1 is not allocated. Since packet 2 contains the time slice allocated for connection 4-1, instead of sending it after packet 1-1 in arrival order, it delays the sending of packet 2 by a time t until a sufficient amount of idle time is found on the outgoing channel . In the example of FIG. 3 , a first sufficiently long idle time period is found on channel B after time B-t7. Time slice 1' is allocated from channel B to the connection of channel 6. Even though all other connections have reclaimed their allocated time slices, packets arriving for the connection on channel 6 will still be forwarded to channel B starting at time B-t7. In this way, the transmission delay can be kept uniform. The delay of the first packet is preferably used, eg to optimize the routing of the packet and to determine its length.

In Fig. 3, the network node has received a second packet addressed to the data transport connection 1-1 during the same repetition period. The groupings are identified by their hatching in the figure. However, instead of sending the packet as soon as it can be sent, i.e. on the channel at time B-t8, it is better to send it only at time A-t1 of the next repetition, i.e. in the time slice allocated for the connection , preferably sent on the same channel. In this way, packets can be forwarded at equal intervals, thereby eliminating variations in transmission delays and making such variations the least problematic. The use of the same channel ensures that packets travel equal distances along equally congested routes. In this way, the influence of the channel variation on the variation of the transmission delay can be eliminated. If the higher rate channel is divided into multiple channels, then the unit of these channels can be viewed as a single channel with multiple layers of time slices. Packets may also be transmitted on any channel, or to any number of channels, as long as the interval between the transmission start times of successive packets destined for the same connection is equal to the repetition period R length.

Although Fig. 3 illustrates the allocation of time slices according to the present invention using the call packet lengths of the TETRA system shown in Fig. 2, it is evident in Fig. 3 that the present invention does not in any way limit the length. If the identifier in the packet contains packet size information and connection information, then the packet size is preferably obtained by examining the identifier. Information about the size of the packet can also be obtained from the information contained in the identifier. For example, in TETRA systems it is sufficient information that the connection is eg a speech connection and that channel coded speech is to be transmitted over it. The packet size can also be checked by delaying the first packet so that it is fully received before starting to send. The packet size can also be checked by sending a control packet to the network node during call setup, which indicates that a packet of a certain length will be sent on the connection.

If there is no spare capacity, resources can be allocated in a number of different ways, which depend at least in part on the resource allocation algorithm of the data transmission system employing the present invention. The invention allows optimizing the transmission capacity, for example by passing time slices of connections to provide the required time slices. The time slices are preferably transmitted in such a way that the time slices are shifted to the same time in a repeat period of another channel, or to an earlier time in a repeat period on the same channel. However, these options are not always feasible. In terms of the quality of the connection, it is important that this transmission be as little as possible. Even if possible, the allocation of a one-time delivery slot is not sent back, but the packet is delivered in a new slot. When allocating resources, an attempt can be made to assign different priorities to different connections, for example such that the time slice allocated for a voice connection is only transmitted if the problem cannot be solved by transmitting the time slice allocated for a data connection. Time has less impact on data connection quality.

In systems according to the TETRA standard, resources are allocated in order of priority and time. A priority is defined in the database for each user and group, which priority is used to allocate resources that have become free, or to terminate calls. For example, an emergency call may terminate all other connections except other emergency calls. If there are no free resources during call setup, the call is queued. If resources are available, they are assigned to the call in order of priority. Following the standard resource allocation principle can be well combined with the transmission capacity optimization brought by the present invention, so as to generate a resource allocation algorithm that provides the best transmission capacity.

Figure 4 is a block diagram illustrating a network node of the invention in a first preferred embodiment. For ease of illustration, the operation of the network node is the same as that described above in connection with FIG. 3 . According to Figure 3, 6 incoming channels lead to the network node and two outgoing channels leave the node. The only restriction on the number of outgoing and incoming channels of a network node according to the invention is that the node has at least one incoming and one outgoing channel. The network node includes a clock CLO for synchronization and timing of operations. The clock is connected to a timer TC, which is cleared at intervals of a repeating period, and which timer indicates the current point in the repeating period at which packets are sent at regular intervals. The timer enables the timing of transmission of packets from network nodes to be correct. The network node also includes an exchange terminal ET for sending and receiving packets from other network nodes or other networks. The network node also includes an maintenance unit OMU, which connects the network node to the network management system. The maintenance unit OMU sets the duration of the repetition period in the network nodes according to the command received from the network management system. The network node also includes a call control unit CCU whose functions are call control and resource management. The call control unit CCU identifies the connection associated with the received packet; checks whether the received packet is the first packet received, or if the connection has been allocated a time slice in which packets can be sent; assigns a time slice to the connection, via the switch The terminal ET controls the transmission of the packets so that the packets are sent in the time slice allocated for the connection, and the packets of the same connection are sent at intervals of a repeating period. In order for the call control unit CCU to be able to function successfully, the unit may for example maintain a table T on the connections it is forwarding. This information can also be stored in other ways. Table T contains information about the connections to which the time slices have been allocated, ie the information of call 3 in the first preferred embodiment, and the moment 4 at which the sending of this call was started. The start time 4 is the time in the repetition period indicated by the timer, at which time packets start to be sent, that is, the time at which time slices in the repetition period start to be reserved. Furthermore, the length of the reserved time slice must be discovered. In this example, the problem is solved by maintaining column 5 in table T, in which the length of the reserved time slice is recorded in bits, from which the length of the time slice can be easily calculated given the transmission rate Duration. The length can also be expressed in seconds. The length of the reserved time slice can also be indicated by indicating the end time of the time slice. This information is the minimum required for the call control unit CCU. This information allows the CCU to allocate time slices, identify connections that have been allocated time slices, and time transmissions correctly.

Table T preferably contains, besides the minimum information, other information, such as information on the packet 6 sent to the connection and, if there are several outgoing channels in the network node, the outgoing channel 7 on which the packet is being sent. Table T can contain both types of information, neither, or only one type of information. Outgoing channel information ensures that packets are actually sent on the same outbound channel, thereby ensuring that transmission delays do not vary by channel.

Figure 4 shows an alternative way of maintaining information about packets sent to the connection. Column 'Sent' 6 in Table T indicates whether the packet has been sent in the previous time slot, as will be described below in conjunction with FIG. 6 . There are other options: for example, this column could contain information about all packets sent, and information about when the packets were sent. The solution shown in Figure 4 can avoid unnecessary capacity allocation. Unlike other schemes, this scheme cannot be used to collect billing information.

In a first preferred embodiment, the call control unit CCU comprises a buffer BUF for buffering packets. The buffer BUF can also be located in other units in the network node, or it can be a separate module. In other embodiments, no separate buffer is needed at all.

The flow diagram of Fig. 5 illustrates a first preferred embodiment of the invention in which paging packets are separated at a network node from control packets containing data transfer to signaling or control information. A separate queue is maintained in the buffers of the network nodes for control packets. In the network node of Fig. 4, the call control unit comprises a buffer. The buffer can also be located in other units in the network node, or it can be a separate module. Call packets can contain voice or data. If the network node receives 20 a new packet, the network node checks 21 whether the packet is a control packet based on identification data contained in the packet. If it is a control packet, it is checked whether a time slice is allocated 23 for the connection associated with it. If a time slice has been allocated for the connection, the start time of packet transmission, ie the output time, is sought 25 . Depending on the start time, the packet either waits for the output time in the buffer, or starts sending the packet immediately. If no time slots are assigned, the call is a new call to which 24 time slots are allocated in the manner described above in connection with FIG. 3 . If a time slice is allocated, then the output time of the packet is determined 25 and the packet is either inserted into a buffer to wait for the output time, or transmission starts immediately.

If the control packets and call packets are not separated, only some steps of the method presented in Fig. 5 are performed. If a new packet is received, it is checked whether a time slice is allocated 23 for the connection associated with this packet. If so, the output time is sought 25 for the packet, and the packet is either inserted into a buffer to wait for the output time, or transmission begins immediately. If no time slice is allocated, then 24 time slices are allocated to determine the starting moment of 25 sending, and the method is the same as that described above.

FIG. 6 illustrates the operation of the transmitting side of the network node according to the invention in a first preferred embodiment, wherein the network node is the network node shown in FIG. 4 . In the case of FIG. 6, it is waiting 26 for the transmission time of the next paging packet. When that moment is reached, it is checked whether there are 27 packets waiting in the buffer. If there is a packet in the buffer associated with the call, the packet is sent, marking 28 a P in the 'sent' column of the table maintained by the network node according to the scheme described in connection with Figure 4, to indicate that the packet has been sent. After sending the packet, calculate 29 the remaining time before the next paging packet is sent, ie the sending moment of the next paging packet. The queue of control packets is then preferably monitored 30 for such a first control message, which is short enough to have sufficient time to send. If one is found 31, then the control message is sent 32.

After sending the control message, it is checked 33 whether it is the sending time of the next paging packet. If yes, the process returns to step 27 to check if any packets are waiting in the buffer. If the transmission time of the next paging packet is not approaching yet, the process returns to step 29, and the interval until the transmission time of the next paging packet is calculated. If no sufficiently short control message is found 31, then wait 26 for the sending moment of the next paging packet. It is also possible not to search for a sufficiently short control message, but to select the control message to be sent in the order of arrival. Therefore, instead of checking whether a sufficiently short control message can be found as before, it is checked whether there is enough time to send the next control message. If there is not enough time, the message returns to the queue, waiting for the sending moment of the next call packet. If call packets have priority, they are inserted into the queue in priority order, and even control packets with the highest priority can be interpreted as call packets, allocating separate time slices to them.

If there is no 27 next call packet, check 34 whether there is an E in the 'send' column of the table according to the scheme given in conjunction with Figure 4. If there is E in this column, it means that there is no grouping in the previous time, and the unnecessary time slice allocated for this call is canceled from the table of Fig. 4, that is, the time slice allocated by 35 is reclaimed. Thereafter, the procedure returns to step 29, and the interval until the transmission timing of the next paging packet is calculated. If the check carried out in step 34 shows that there is P in the 'sent' column of the table given in Figure 4, then E is marked 36 in the 'sent' column, and the process proceeds to step 29 to calculate the time to send the next call packet interval. Cancellation of time slice allocation, ie reclaiming of allocated time slices, may also be due to a call termination message indicating that the call has been terminated. In this way, the above-mentioned reclaiming of the allocation of time slices after two consecutive 'no packet' indications improves the efficiency of transmission capacity, as unused capacity becomes available here. Furthermore, there is no need to send a separate call termination message, thus reducing the amount of signaling. In fact, the absence of a packet can also be detected in many other ways, for example by comparing the interval to the previous packet sent, as described above in connection with FIG. 4 . The number of packets that must be missing before reclaiming the allocated time slice may also be other than 2, but this number must be determined in advance and the network nodes must function properly. If a packet associated with a call is received for forwarding after the time slot allotted for the call has been reclaimed, the call can be considered a new call for which a new time slot will be assigned.

If the control packets and call packets are not separated, only some steps of the method presented in Fig. 6 are performed. Call packets here refer to all packets that need to be sent. The initial situation is to wait 26 for the sending moment of the next paging packet. If the moment arrives, it is checked whether there are 27 packets waiting in the buffer. If there is a packet in the buffer associated with the connection, the packet is sent, marking 28 a P in the 'send' column of the table maintained by the network node according to the scheme described in connection with Figure 4, to indicate that the packet has been sent, The process then proceeds to step 26 to wait for the transmission moment of the next paging packet. If there is no next call packet, check 34 whether there is an E in the 'Send' column of the table according to the scheme given in connection with Figure 4. If there is E in this column, it means that there was no grouping in the previous time, and then cancel the unnecessary time slice allocation for this connection from the table in Figure 4, that is, reclaim 35 allocated time slices. Thereafter, the process returns to step 26, waiting for the transmission moment of the next paging packet. If the inspection carried out in the step 34 shows that there is P in the 'sending' column of the table given in Figure 4, then mark E in the 'sending' column, the process proceeds to step 26, waiting for the sending moment of the next call packet.

With regard to the different processing of packets based on the information content of the packets indicated by the identification information in Figures 5 and 6 and related descriptions, it should be pointed out that similar divisions can also be achieved by separating packets containing speech from other packets, so that the speech packet The processing method is the same as that of the call packets in Figures 5 and 6, that is, only the voice packets are allocated time slices. Control packets and paging packets containing data are handled in the same manner as paging packets given in the description in connection with Figures 5 and 6, ie these packets are sent between voice packets as long as there is sufficient time between them. These packets can also be divided into control, voice, and data packets, with voice and data packets being allocated time slices with higher priority than control packets. It is also possible to roughly divide the packets into control packets and call packets, but in the case of congestion, the call packets are divided into voice packets and data packets, so the different types of packets can be handled differently. If real-time images are transmitted in the network, image packets are handled in the same way as voice packets.

The figures and the associated description are only intended to illustrate the invention and its application in mobile systems. It will be apparent to those skilled in the art that the present invention can be varied and modified in many ways without departing from the scope and spirit of the invention as disclosed in the appended claims.

Claims (11)

1. method for packet switched data transmission, wherein the grouping that will receive is sent to the destination address of grouping, it is characterized in that may further comprise the steps:

Determine from the repetition period, to distribute the timeslice of different length for connection more than sending the long repetition period of single grouping required time;

Determine the starting point of repetition period;

Safeguard the information of current point in the repetition period;

The timeslice information that maintenance distributed from the repetition period, and the link information that is associated with it;

Receive grouping (20);

According to the identification data of grouping, identification and being connected and destination address that grouping is associated;

Check whether from the repetition period, distributed timeslice for the connection of dividing into groups; And

If when having arrived the starting point of the timeslice of distributing in the repetition period, begin to send grouping so;

If not,

-determine that grouping sends duration;

-search sends the same long free time section of duration with this at least in the repetition period;

-free time from the repetition period is in the section, sends the same long timeslice of duration for the connection of grouping distributes at least with grouping; And

-when having arrived the starting point of the timeslice of distributing in the repetition period, begin to send grouping.

2. according to the method for claim 1, it is characterized in that, further comprising the steps of:

If such two continuous repetition periods are arranged, then regain to this and connect the timeslice (35) of distributing: during this two repetition periods, the grouping that joins with this join dependency need not send in timeslice, and

If after regaining the timeslice of distributing, receive the grouping that joins with this join dependency, then distribute a new timeslice for this connection.

3. according to the method for claim 1 or 2, it is characterized in that, further comprising the steps of:

For each available channel of going out is determined repetition period and starting point thereof separately,

Select the go out channel of such channel as new connection: the fastest discovery is fit to the timeslice of distribution on this channel,

On the same channel of going out, send the grouping that joins with identical join dependency, and

Safeguard the timeslice information of from the repetition period, distributing in the specific mode of channel, and the link information that is associated with it.

4. according to claim 1,2 or 3 method, it is characterized in that,

According to the identification data that is included in the grouping, grouping is divided into the control grouping of calling out grouping and comprising signaling information,

Only be to distribute timeslice with calling out being connected of dividing into groups to be associated,

In the timeslice of distributing, send and call out grouping, and

In the repetition period, send the control grouping on such point: on these aspects, control can be sent fully divide into groups before the timeslice that the next one distributes.

5. according to claim 1,2 or 3 method, it is characterized in that,

According to the identification data that is included in the grouping, grouping is divided into packets of voice and other grouping that comprises voice,

Only for distributing timeslice with being connected of being associated of packets of voice,

In the timeslice of distributing, send packets of voice, and

In the repetition period, send other grouping on such point: on these aspects, can before the timeslice of next one distribution, send other grouping fully.

6. according to method any in the aforementioned claim, be used for the transmission network of mobile telephone system, it is characterized in that, the length of repetition period is defined as sends the time span that frame is required on the wireless path.

7. packet switched data transmission network of network node, have at least one to enter channel towards this network node, have the channel sources of going out at least from this network node, this node comprises the dispensing device that sends grouping to destination address, it is characterized in that this node also comprises

Setting device (OMU) is arranged to be longer than the single transmission packets time with the persistence length of repetition period;

At least one timer (TC), in order to the current point in the indication repetition period, with the timed sending grouping, this timer is with the interval zero clearing of a repetition period;

Recognition device (CCU) is used to discern and being connected that the grouping that receives is associated, and detects network node whether received grouping with this join dependency connection before described grouping, and perhaps whether this grouping is and first dividing into groups that this join dependency joins;

Distributor (CCU), be used for sending the same long free time sheet of duration with grouping at least in the repetition period search, divide into groups in response to receiving with first of this join dependency connection, from this sheet, send the same long timeslice of duration with grouping for distributing at least with connection that this grouping is associated free time, the starting point of distributing timeslice is by the timer indication of determining with transmission zero hour of the grouping of this join dependency connection, and

Control device (CCU) is used for for sending grouping with the timeslice of distributing that is connected that this grouping is associated to destination address, makes that next sends the grouping with same join dependency connection group by group from network node one with the interval of a repetition period.

8. according to the network node of claim 7, it is characterized in that, also comprise tape deck (7), the timeslice that is used to the grouping of each linkage record transmission and does not have to divide into groups, and, regain the timeslice of distributing in response to last two sheets continuous time that do not have grouping of same connection.

9. according to the network node of claim 7 or 8, comprise that also at least the second channel of going out is characterized in that,

Each channel has the different repetition periods;

Distributor (CCU) is each channel allocation timeslice, and

Control device (CCU) sends the grouping with same join dependency connection on the same channel of going out.

10. according to claim 7,8 or 9 network node, it is characterized in that,

Network node is arranged in the transmission network of mobile telephone system,

Send a required time span of frame on repetition period and the wireless path and equate, and

(PL1 PL2), determines the length of the grouping (PA) that needs send according to the block length that sends on the wireless path in the mobile telephone system.

11. the network node according to claim 10 is characterized in that,

Identity device (CCU) separate calls grouping and comprise the control grouping of control information,

Distributor (CCU) only distributes timeslice after receiving the first calling grouping, and

Control device (CCU) sends the control grouping in the interval that can send the control grouping fully between the timeslice of distributing.

Images