JPH022756A - Packet transfer system - Google Patents

Abstract

PURPOSE:To economically and efficiently perform packet transfer by setting information to instruct the possibility of deletion of a packet when the packet encounters congestion in a switching machine and the information to instruct the necessity of re-transmission when it encounters a transmission path error. CONSTITUTION:The information from a TV camera 3 and a mike 4 as terminals, for example, TV conference terminals are assembled as the packet at a packet disassembling device 5, and is transferred to a packet switching machine 7, and the switching machine 7 transfers the packet after dividing to constant length. On the packet, a deletion classification field to instruct the possibility of the deletion of the packet in the case of encountering to the congestion in the switching machine and a re-transmission classification field to instruct the necessity of re-transmission in the case of encountering to the transmission path error are set. Furthermore, the transfer of the packet from the switching machine 7 to another switching machine is performed with an 1 frame when the re-transmission is needed, and with a UI frame when it is not needed, and the switching machine 8 receiving the packet controls the deletion when the congestion is generated in an output route based on the information of the possibility of the deletion. In such a way, it is possible to perform the packet transfer economically and efficiently.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a packet transfer system for a packet switching network for multimedia such as voice, data, and video.

[Conventional technology]

Conventional packet-switched networks are mostly intended only for packet transfer between data terminals. Packet transfer between data terminals does not require much real-time performance, but is characterized by requiring high reliability against errors. In a conventional packet switching network that satisfies this condition, error recovery of X, 25 level 2 is performed between adjacent exchanges, and error recovery of X, 25 level 3 is performed between originating and terminating exchanges.

Note that X, 25 is the recommendation number of the CCITT (International Telegraph and Telephone Consultative Committee) recommendation, which specifies (packet-type terminal interface) and defines basic specifications for the following three levels.

L/1: Physical level interface Level 2: Link level protocol Level 3: Packet level protocol On the other hand, in recent packet switching networks, the transfer speed is high, and the target media is from the data terminal to It is expanding not only to data, but also to audio and video information. A comparison of the characteristics of these media with data reveals the following.

In order to maintain the naturalness of the conversation, voice needs to be delivered to the other party with a shorter delay time than packet transfer between data terminals. That is, since the delay time for transmitting the audio packet to the other party must be at least 150 milliseconds or less, this means that if the audio packet is erroneous, it cannot be reproduced. Also, unlike data, voice information has redundancy, so errors of about a few percent can be tolerated.

Video information such as video conferences requires real-time performance, short delay time due to the need to synchronize video and audio, and fewer errors due to differential encoding such as interframe prediction. is necessary.

[Problem to be solved by the invention] As described above, the requirements for networks regarding delay time and errors differ depending on the media, so it is difficult to mix multimedia information in a conventional switching network that mainly targets packet transfer between data terminals. It would have been inappropriate to transfer the information.

By making it possible to specify whether or not to discard or not to retransmit packets when transferring packets, the present invention provides a means for transferring packets according to the media for media with different properties, thereby providing economical and efficient multimedia packets. The purpose is to realize transfer.

[Means to solve the problem]

The present invention sets information in a packet to instruct whether or not to discard the packet if the packet encounters switching congestion, and information to instruct whether or not to retransmit if the packet encounters a transmission path error. When transferring packets between exchanges, packets that need to be retransmitted are sent as ■ frames, and packets that cannot be retransmitted are sent as UI frames. If the route is congested, discard control is performed based on information indicating whether discard is possible, and this is the most important feature.

[Effect]

The means for solving the above problem works as follows.

Figure 2 shows the frame structure for transferring packets, and the frame consists of flag sequences 31, 36.
, address, l+I [Il, information, and frame check sequence fields 32 to 35. Also, Figure 3 shows the configuration of the control field 33 in Figure 2!
It shows frames and UI frames. ■
Sequence numbers N(S) and N(R) indicating the frame transmission/reception order are set in the control field of the frame.

If the frame is incorrect these sequence numbers N(S)
, N(R) performs packet frame retransmission control. On the other hand, since the control field of the UI frame does not have a sequence number as shown in the figure, the frame is not retransmitted even if an error occurs. Therefore, either an I frame or a UI frame is selected and transmitted depending on whether retransmission is necessary.

Also, if the outgoing route corresponding to the packet destination is congested,
The transmission waiting queue corresponding to that output route becomes longer, and the delay time required for packet transfer also becomes longer. Therefore, if a packet that does not satisfy real-time performance, such as a voice packet, can be discarded, specifying that it can be discarded allows the packet to be discarded when congestion occurs.

〔Example〕

FIG. 1 is a block diagram showing one embodiment of the present invention. In the same figure, 1 and 11 are PBX, and 2.12 is PBX packet assembly and disassembly! (PAD), 3.13 is a TV camera and monitor, 4.14 is a microphone and speaker, 5 and 15 are packet assembly and disassembly equipment for video conferencing (P
AD), 6.16 is a packet data terminal, and 7, 8.9 is a packet switch.

Further, FIG. 4 shows the format of the packet used in the present invention, that is, the contents of the information field 34 in FIG. 2, where 21 is a retransmission type field indicating whether retransmission is necessary;
Reference numeral 22 indicates a discard type field indicating whether or not discard is possible. A detailed explanation will be given below based on these figures.

In Pakeno 14M, information from 3.4 as a terminal, for example, a video conference terminal, is sent to the packet assembly and disassembly device (PAD).
The packets are assembled into packets in step 5 and transferred to packet switch 7. The length of the input packet to the packet switch 7 is selected appropriately from among the user-network interface conditions. This also applies to input data from the packet data terminal 6 and the like. When packets input from these terminals are transferred within the packet switching network with the same length, if the input packet is a long packet such as 4096 bytes, there will be a delay time compared to a packet requesting short delay transfer. It makes the . For this reason, the present invention is directed to a switching network in which if an input packet to a packet switch is longer than a certain length (for example, longer than 128 bytes), the packet switch 7 divides it into pieces of a certain length and transfers the packet.

In such a switching network, for example, voice information from the microphone 4 as a video conference terminal requires real-time quality, so there is no need to retransmit it, and it is possible to discard it in the event of congestion. Further, it is assumed that the information from the television camera 3 cannot be retransmitted, cannot be discarded, and should be transferred. In this case, for the information from the microphone, for example, set the value of the Q bit of X, 25 level 3 to "1°" and transfer it to the packet switch 7 in the UI frame, and for the information from the TV camera 3. The video conference packet assembly and disassembly device 5 is configured to set the value of the Q bit of x, 25 level 3 to "0" and transfer it to the packet switch 7 in a UI frame.
Here, the X, 25 level 3 Q bit is used to instruct the packet switch from the terminal as to whether or not it can be discarded, but it is also possible to designate it in an appropriate field of the user data, etc. transmission. Further, information etc. from a data terminal is usually retransmitted and cannot be discarded, but in this case, the Q bit is set as OII and is configured to be transferred in 1 frame.

When the packet switch 7 receives the above information from the terminal, the retransmission type field (21 in Figure 4) of the interstation header used within the network indicates that the l frame is a UI frame, that is, if there is an error on the transmission path, it will be retransmitted. Set the type of information, whether it is necessary information or unnecessary information. Furthermore, based on the value of the Q bit, if the exchange that received the packet becomes congested, the discard type field (22 in Figure 4), which indicates whether discarding the packet is allowed, is set to indicate whether discard is allowed or not. Set. Furthermore, when the packet switch 7 transfers a received packet to another packet switch 8, it selects and transfers either an l frame or a UI frame depending on the value of the retransmission type field, so when a transmission path error occurs, an l frame is used. Only the transmitted frame is retransmitted by the level 2 retransmission means of X,25, and the UIl
Frames that have been sent will not be retransmitted. Furthermore, when the packet switch 8 becomes congested and the queue waiting to be sent to the transmission line becomes long, the congestion state is recovered by discarding packets for which discardable is specified in the discard type field. Regarding the packet discard method, all discardable frames connected to the queue may be discarded, or some discardable frames may be discarded.

〔Effect of the invention〕

As explained above, according to the present invention, voice, data,
It becomes possible to transfer multimedia information such as video according to the characteristics of the media, and economical and rational multimedia packet transfer becomes possible.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is an explanatory diagram showing an example of the structure of a frame used in the present invention, and FIG. 3 is an example of the structure of the control field in FIG. FIG. 4 is an explanatory diagram showing the contents of the information fields in FIG. 2. Explanation of symbols

Claims (1)

[Claims] 1) In a packet switching network in which packets from a terminal are divided into pieces of a certain length or less and transferred within the network, the packet is provided with information to instruct whether or not to discard it if the packet encounters switching congestion. , and information for instructing whether retransmission is necessary when a transmission path error is encountered, and when transferring packets between exchanges, an I frame is used if retransmission is necessary, and a UI frame is used if retransmission is not necessary. A packet transfer method characterized in that, in an exchange that transmits a packet and receives the packet, if an outgoing route to the destination of the packet is congested, discard control is performed based on the discard permission/prohibition information.