JP2003304273A - Device, program and method for repeating packet - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a technique for more surely realizing packet retransmission between repetition nodes even in the case of a packet communication network in which the communication path is changed at any time. <P>SOLUTION: A packet repeater is provided with a repetition part, a temporary storage part, a retransmission requesting part and a retransmission part. The repetition part transmits a packet received from the previous stage via the packet communication network to the following stage upon adding address information of it (repetition node) to the packet. The temporary storage part temporariry stores the packet to be repeated by the repetition part. The retransmission requesting part reads the address information indicating the previous stage from packets before or after packet loss when the packet loss is caused in the packet to be repeated by the repetition part. Furthermore, the retransmission requesting part transmits a retransmission request to the previous stage indicated by the address information. The retransmission part retransmits the packet temporarily stored in the temporary storage part when the retransmission request is received from the outside. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a packet relay device suitable for a packet communication network whose communication route is changed such as the Internet (when a communication route selecting function is provided,
Also referred to as a packet switching device) and a packet relay method. The present invention relates to a packet relay program for causing a computer having a communication function to function as a packet relay device.

[0002]

2. Description of the Related Art Conventionally, as a method of correcting an error in packet communication, "a method of performing error correction on the receiving side" and "a method of retransmitting a packet between a start point and an end point of a communication path".
Is known.

[0003]

This error correction method has a problem in that the coding efficiency is reduced by the addition of the error correction code to the communication data. Further, there has been a problem that error correction cannot be performed for a data error in a unit of a certain amount.

Further, the method of retransmitting a packet between the start point and the end point of a communication route has a problem in that the packet retransmitting route tends to be long and it takes time to retransmit the packet. Especially for real-time media (video, audio, etc.), such a packet retransmission method is not suitable because the allowable delay time is short.

Further, since the packet is retransmitted between the start point and the end point of the communication path, there is a problem that the traffic of the packet communication network tends to increase. In particular, in physically large networks such as the Internet and in multicast communication with many receivers, retransmission requests and retransmission packets occur frequently, and each time, each communication path is temporarily occupied over the entire length. . Therefore, there has been a problem that the communication capacity of the packet communication network is likely to be reduced in general.

Prior to the Internet, a communication system called a high level communication control procedure (HDLC) is known. In this HDLC, since the communication path is fixed, it is easy to realize packet retransmission between fixed relay nodes. However,
In a communication network such as the Internet, even if the start point and the end point of the communication route are fixed, the relay nodes on the way are changed at any time, so that it is very difficult to realize packet retransmission between the relay nodes.

In view of the above-mentioned conventional problems, it is an object of the present invention to more reliably realize packet retransmission between relay nodes even in a packet communication network whose communication route is changed at any time.

[0008]

The present invention will be described below. <Claim 1> A packet relay device according to claim 1 is a packet relay device arranged in a relay node of a packet communication network, and includes a relay unit, a temporary storage unit, a retransmission request unit, and a retransmission unit. The relay unit adds the address information of its own (relay node) to the packet received from the former stage via the packet communication network, and then relays the packet to the latter stage. The temporary storage unit temporarily stores the packet relayed by the relay unit. When a packet loss occurs in the packet relayed by the relay unit, the retransmission request unit reads the "address information added in the preceding stage" from the packet before or after the packet loss. Then, the resend request unit sends a resend request to the preceding stage indicated by this address information. Upon receiving a retransmission request from the outside, the retransmission unit retransmits the packet temporarily stored in the temporary storage unit. << Claim 2 >> The packet relay device according to claim 2 is
In the packet relay device described in the above item 1, the relay unit detects a packet loss from a mismatch of information indicating a transmission order of consecutive packets (hereinafter referred to as “order information”). << Claim 3 >> The packet relay device of claim 3 is
In the packet relay device described in the item [1], the relay unit corrects the inconsistency of the order information with respect to the subsequent stage. << Claim 4 >> The packet relay device of claim 4 is
In the packet relay device according to any one of claims 3 to 3, the temporary storage unit releases the storage area of the packet after the allowable delay of the packet elapses. << Claim 5 >> The packet relay device according to claim 5 is
In the packet relay device according to any one of claims 4 to 4, the retransmission unit updates the retransmission counter added to the packet for each retransmission. Then, the retransmission unit stops the retransmission when the number of retransmissions or the retransmission time indicated by the retransmission counter reaches or exceeds a predetermined value. <Claim 6> The packet relay program according to claim 6 includes a computer having a packet communication function, the relay section, the temporary storage section, the resend request section, and the resend section according to any one of claims 1 to 5. The feature is that it functions as a section. <Claim 7> The packet relay method according to claim 7 is a packet relay method implemented in a relay node of a packet communication network, and includes a relay step, a temporary storage step, a retransmission request step, and a retransmission step. In this relay step, the address information of the self (relay node) is added to the packet received from the former stage via the packet communication network, and then the packet is relayed to the latter stage. In the temporary storage step, the packet relayed in the relay step is temporarily stored. In the retransmission request step, when packet loss occurs in the packet relayed in the relay step, the "address information added in the preceding stage" is read from the packet before or after the packet loss. Then, in this retransmission request step, a retransmission request is transmitted to the preceding stage indicated by this address information. In the retransmission step, when a retransmission request is received from the outside, the packet temporarily stored in the temporary storage step is retransmitted. << Claim 8 >> The packet relay method of claim 8 is the same as that of claim 7.
In the packet relay method described in (1), in the relay step, a packet loss is detected from a mismatch of information indicating a packet sending order (hereinafter referred to as “order information”). << Claim 9 >> The packet relay method of claim 9 is the same as that of claim 8.
In the packet relay method described in the step of relaying,
Correct the inconsistency of the order information for the latter stage. <Claim 10> The packet relay method according to claim 10 is the packet relay method according to any one of claims 7 to 9, wherein the temporary storage step stores the packet after the allowable delay of the packet has elapsed. Free up space. <Claim 11> The packet relay method according to claim 11 is the packet relay method according to any one of claims 7 to 10, wherein in the retransmission step, a retransmission counter added to the packet is updated for each retransmission. To do. Then, in this retransmission step, the retransmission is stopped when the number of retransmissions or the retransmission time indicated by the retransmission counter reaches or exceeds a predetermined value.

[0009]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings.

[Description of Configuration of Packet Relay Device] FIG.
It is a figure which shows the structure of the packet relay apparatus 11. As shown in FIG. 1, the packet relay device 11 includes a relay unit 12 that relays a packet, a temporary storage unit 13 that stores a relayed packet, and a retransmission request unit 1 that issues a packet retransmission request.
4 and a resending unit 15 that resends the packet. Note that these constituent requirements may be implemented as hardware. Further, by executing the packet relay program in a computer having a packet communication function, it is also possible to realize these constituent elements by software.

[Description of Packet Communication Network] FIG. 2 is a diagram showing a state of the packet communication network. In this packet communication network, a communication path starting from a packet transmission source 21 and ending at a packet destination 22 is set at any time. A plurality of relay nodes 23 are provided on the communication path. In all or part of such a relay node 23,
The packet relay device 11 shown in FIG. 1 is arranged.

[Explanation of Data Structure of Packet] FIG.
3 is a diagram showing a data structure of a packet relayed by the packet relay device 11. FIG. As shown in FIG. 3, the packet is roughly divided into packet data and a header area. The following information is set in this header area.

(1) Destination address of packet (for example,
(IP address) (2) Source address of packet (eg IP address) (3) Allowable delay of packet (generally added at the source, eg deadline expressed in absolute time, (It may be a delay time represented by a relative time) (4) Sequence information (for example, a serial number indicating the sending order of packets) (5) Retransmission counter (initial value is the maximum number of retransmissions allowed) (6) Address of relay node Information (eg IP address)

[Description of Operation of Packet Relay Device 11] FIG. 4
~ FIG. 7 is a flowchart illustrating the operation of the packet relay device 11. Hereinafter, the operation of the packet relay device 11 will be described according to the step numbers shown in FIGS.

Step S1: The relay unit 12 in the packet relay device 11 determines whether or not a retransmission request (NAK packet or the like) has been received. Here, if there is no resend request,
The relay unit 12 shifts the operation to step S8. On the other hand, when the retransmission request is received, the relay unit 12 performs step S2.
Move to.

Step S2: The relay unit 12 uses the retransmission unit 1
The message of "retransmission request reception" is sent to 5. The resending unit 15 responds to the message “retransmission request received” by the message handler (steps S3 to S) shown in FIG.
Execute 7).

Step S3: The retransmitting section 15 specifies a packet specified in the retransmission request (hereinafter referred to as "retransmission packet").
However, it is determined whether or not it exists in the temporary storage unit 13. When a retransmission packet does not exist in the temporary storage unit 13 due to the reason that an allowable delay (described later) has passed, the retransmission unit 15
The packet retransmission operation is abandoned, the message handler is terminated, and the operation proceeds to step S8. On the other hand, when the retransmission packet exists in the temporary storage unit 13, the retransmission unit 15 shifts the operation to step S4.

Step S4: The resending unit 15 decrements the resending counter of the resending packet by one. Note that this retransmission counter is clearly different from the information such as the time to live of the packet (TIME TO LIVE) in that it counts the retransmission operation.

Step S5: The retransmitting section 15 determines whether or not this retransmission counter has a negative value. Here, when the retransmission counter has a negative value, the retransmission unit 15 moves the operation to step S6. On the other hand, if the value of the retransmission counter is zero or more, the retransmission unit 15 shifts the operation to step S7.

Step S6: In the retransmission counter, the allowable upper limit of the number of retransmissions is initialized. The fact that this retransmission counter has a negative value means that the number of past retransmissions has exceeded the allowable upper limit. In this case, the resending unit 15
Cancels the further retransmission operation and releases the storage area of the retransmission packet. After such an operation, the resending unit 15
The message handler is terminated and the operation proceeds to step S8.

Step S7: The retransmitting section 15 corrects the inconsistency in the order information of the retransmitted packet and retransmits the retransmitted packet to the subsequent stage. It should be noted that, in order to prevent the packet loss from occurring again, the route change may be carried out triggered by this retransmission operation. After such an operation, the resending unit 15 terminates the message handler and returns to step S8.
Move to.

Step S8: The relay unit 12 monitors the arrival of packets. If no packet arrives at this point, the relay unit 12 returns to step S1. on the other hand,
When the arrival of the packet is detected, the relay unit 12 shifts the operation to step S9.

Step S9: The relay section 12 sends a message of “packet arrival” to the temporary storage section 13.
In response to this "packet arrival" message, the temporary storage unit 13 executes the message handler (steps S10 and S11) shown in FIG.

Step S10: The temporary storage unit 13 secures a storage area according to the size of the incoming packet,
The incoming packet is temporarily stored in this storage area.

Step S11: Next, the temporary storage unit 13
Searches the stored packets for which the allowable delay has elapsed. Here, when a packet for which the allowable delay has elapsed is found, the temporary storage unit 13 releases the storage area of the packet. After such an operation, the temporary storage unit 13 terminates the message handler and shifts the operation to step S12.

Step S12: The relay section 12 reads out the order information from the header area of the incoming packet. further,
The relay unit 12 inquires of the temporary storage unit 13 or the like to search for a immediately preceding packet (hereinafter referred to as “previous packet”) in a packet group having the same source and the same destination as the incoming packet.
The relay unit 12 compares the order information of the immediately preceding packet with the order information of the incoming packet to determine whether or not the order information of the incoming packet is inconsistent. This mismatch determination operation will be specifically described with reference to FIG.
In FIG. 8, squares represent individual packets, and the numbers in the squares represent order information added to each packet. As shown in FIG. 8, a packet having sequence information of "1", "2", "4", "5", "6", "7" arrives from the preceding relay node A to the relay node M. To do. This order information is information indicating the sending order of packets, and has a predetermined change rule (incremented by 1 here). Relay node M
In the light of this change rule, the relay section 12 arranged in
A mismatch is detected between the order information “2” and the order information “4”. The relay unit 12 specifically determines the loss of the packet having the order information “3” based on such a mismatch.

Step S13: When the relay section 12 detects the inconsistency in the order information of the incoming packet, the step S1
The operation shifts to 4. On the other hand, when no inconsistency is detected in the order information of the incoming packet, the relay unit 12 determines in step S1.
The operation shifts to 7.

Step S14: The relay unit 12 sends a message "packet loss occurred" to the retransmission request unit 14. In response to this "packet loss occurrence" message, the resend request unit 14 executes the message handler (steps S15 and S16) shown in FIG.

Step S15: The resend request unit 14 selects at least one packet received before or after packet loss from the packet group of the same source and the same destination. In this case, it is simple and preferable to select two packets (in the case of FIG. 8, a packet having the order information “2” and a packet having the order information “4”) in which the mismatch of the order information is detected. The retransmission request unit 14 accesses the temporary storage unit 13 and reads the “relay node address information” added in the previous stage from the header area of the selected packet. Regarding the relayed packet, the "relay node address information" is rewritten to its own address information (see step S17). Therefore, it is preferable to take measures such as separately storing the "address information of the relay node" added in the previous stage.

Step S16: When the address information of the relay node matches in the packet received before and after the packet loss, there is a high possibility that the lost packet is temporarily stored in the relay node. In this case, the retransmission request unit 14 transmits a retransmission request (NAK packet or the like) to the relay node indicated by this address information. On the other hand, when the address information of the relay node does not match in the packets before and after the packet loss, it can be determined that the communication route has been changed before and after the packet loss.
In this case, the retransmission request unit 14 determines that both relay nodes (see FIG.
In the case of 1, it is preferable to promptly send a retransmission request to the relay nodes A and B). By such an operation, it becomes possible to reliably receive the packet retransmission service from either the relay node A or the relay node B.

In such a case, it is possible to fully anticipate that the order of arrival of packets is simply changed due to the change of the communication route. That is, as shown in FIG. 11, when the order information “3” and the order information “4” pass through different communication paths, the packet of the order information “4” may happen to arrive first. In order to deal with such a situation, the retransmission request unit 14 waits for the arrival of “packets whose order of arrival is reversed” for a predetermined time so that unnecessary retransmission requests are not transmitted, and then the both sides do not exceed the allowable delay. It is preferable to send a resend request to the relay node.

Incidentally, it is expected that the "relay node address information" does not exist in the header area of the selected packet. In such a case, the resend request unit 14
It is preferable that the address information of the transmission source is read from the header area of the selected packet and a retransmission request (NAK packet or the like) is voluntarily transmitted to the packet transmission source indicated by the address information.

In order to omit such an irregular operation, it is preferable that the transmission source initializes its own address information in "address information of relay node". With this initial setting, the relay node can unconditionally make a retransmission request to the preceding stage indicated by the "relay node address information" without particularly distinguishing the transmission source and the relay node. After such an operation, the resend request unit 14 ends the message handler and moves the operation to step S17.

Step S17: The relay unit 12 accesses the header area of the incoming packet and rewrites the "address information of the relay node" with its own address information.

Step S18: The relay section 12 rewrites the order information of the incoming packet so that the order information matches the immediately preceding packet. This sequence information rewriting operation will be described with reference to FIG. In FIG. 8, a packet having sequence information of "1", "2", "4", "5", "6", "7" arrives from the relay node A at the preceding stage to the relay node M. It is assumed that a packet is transmitted to the subsequent stage (relay node P shown in FIG. 8) with this order information unchanged. In this case, the relay nodes connected in the subsequent stage redundantly detect the inconsistency between the order information “2” and the order information “4”. As a result, useless retransmission requests are generated in a chain from the relay nodes connected in the subsequent stage. Due to such chained retransmission requests, the traffic of the packet communication network is confused. Therefore, FIG.
As shown in, in the relay node M that has transmitted the retransmission request,
The order information of each packet is rewritten and corrected to "1", "2", "3", "4", "5", "6". By correcting the order information in this way, it is possible to hide the inconsistency of the order information from the subsequent stage and avoid a situation in which unnecessary retransmission requests are generated in a chain.

Step S19: The relay section 12 selects a line based on the destination information of the incoming packet and transmits the incoming packet (the above-mentioned rewritten one) to the subsequent stage.
After such an operation, the relay unit 12 returns the operation to step S1 again.

[Verification result of this embodiment] FIG.
Is an experimental network for verifying the effect of reducing the retransmission time. In this experimental network, a relay node 2 to which the present embodiment is applied is provided between a source 21 and a destination 22.
Two 3 are provided and a link 24 is stretched between each node.
Furthermore, in this experimental neckwork, the following communication conditions were set.・ Allowable packet delay is 500 [ms] ・ The packet loss occurrence rate per link is 10% ・ Allowable maximum number of retransmissions is 2 times (theoretical value) and 4 times (the experimental value) 5 is a graph showing a packet discard rate in this experimental network. As shown in FIG. 10, by applying the present embodiment, the packet discard rate is generally improved as compared with the conventional technique (packet retransmission between the start point and the end point).

[Effects of this Embodiment] As described above, in this embodiment, the relay node adds the address information of itself (relay node) to the packet and relays it. Generally, in a packet communication network, even if the communication path is frequently changed, it is rare to change the communication path in units of packets. In reality, the packet communication path is not changed for a predetermined period (for example, in the case of the Internet, the communication path is not changed for several to several tens of seconds,
During that period, the packets pass sequentially through the same relay node).

In this embodiment, attention is paid to the fact that the continuity of the bucket group is maintained for a short time. That is, by referring to the "relay node address information" from the packet before or after the packet loss, the relay node temporarily storing the lost packet can be known with high probability. (Thus, the extent to which the "before" and "after" of this packet loss extends over time can be judged from the period during which the continuity of the packet group is maintained.)

Therefore, in the present embodiment, a retransmission request is sent to the preceding relay node indicated by this "relay node address information". As a result, even in a situation where the communication route is changed from time to time, it is possible to successfully retransmit a packet between relay nodes with a high probability.

In the present embodiment, the packet loss is detected by checking the inconsistency of the order information. With this detection method, it is possible to detect a packet loss during packet relaying without waiting for the reconstruction of the packet group at the packet destination. Therefore, the present detection method is an excellent detection method for detecting packet loss at a relay node on the way.

Further, in the present embodiment, the relay node which has transmitted the retransmission request corrects the inconsistency of the order information and then relays the packet. Therefore, it becomes possible to conceal the inconsistency of the order information from the latter stage and prevent a chained retransmission request from the latter stage.

Further, in the present embodiment, the device for deleting the relayed packet is also devised. Normally, in a relay node, if relayed packets are stored one after another,
The storage area will eventually fill up. Therefore, it is possible to automatically delete old relayed packets.

However, the packet communication network may require highly reliable data communication. In this case, no matter how late it may be, the resend request must be met.

On the other hand, a packet communication network may be used to transmit real-time media. In this case, if it is delayed to the extent of impairing the real-time property, packet retransmission becomes meaningless.

Therefore, in this embodiment, information about the allowable delay is added to the packet, the passage of the allowable delay is determined, and the relayed packet is deleted. Normally, for a packet in which no data error is allowed, the allowable delay is set to be sufficiently long at the transmission source or the like. On the other hand, in the case of real-time media, the allowable delay is set to be short at the transmission source.

Therefore, it becomes possible to properly determine the erase timing of the relayed packet in accordance with the permissible delay for each packet. As a result, it becomes possible to respond to the retransmission request for a period according to the allowable delay for each packet. Further, it is possible to improve the situation in which the relayed packets that have become unnecessary beyond the allowable delay occupy the storage area of the temporary storage unit 13 forever.

Further, in the present embodiment, the number of retransmissions is managed by the retransmission counter added to the packet, and the packet retransmission is stopped when the number of retransmissions exceeds a predetermined value. Therefore, it is possible to prevent a situation in which the retransmission operation is repeated more than necessary for one packet.

Particularly, this retransmission counter is updated by being added to the packet while passing through each relay node. Therefore, the total number of retransmissions in each relay node can be managed by this retransmission counter. Therefore, it becomes possible to monitor the total number of retransmissions by the retransmission counter and stop the packet retransmissions in good timing.

[Supplementary Information of Embodiment] In the above-described embodiment, the information regarding the allowable delay is added to the packet. However, the present invention is not limited to this. For example, based on information indicating the data type of a packet, the type of packet communication (for example, service type or packet priority), the packet relay device 1
1 may determine the allowable delay.

Further, in the above-described embodiment, the relayed packets that have passed the allowable delay are erased at the timing of temporarily storing the latest incoming packet. In this case, since the storage area is occupied and released at the same time, there is an advantage that the storage area is less likely to be full. However, the present invention is not limited to this. For example, the allowable delay may be monitored, and the storage area of the relayed packet may be released immediately after the allowable delay elapses. Furthermore, if this release operation is executed in multithreads, it is possible to release the storage area by tapping the mole. These operations have the advantage that the free space in the storage area can always be secured as much as possible.

Further, for example, the storage area of the relayed packet which has passed the allowable delay may be released at the timing when the free space of the storage area becomes equal to or less than the threshold value. In this case, as long as the storage area has enough space, it is possible to continue to store the relayed packets and meet the request for packet retransmission as much as possible.

In the above-described embodiment, the retransmission counter manages the number of packet retransmissions. However, the present invention is not limited to this. For example, by setting the update width of the retransmission counter to "delay time associated with retransmission" or the like, it becomes possible to manage the retransmission time of the packet. In this case, it is preferable to set the initial value of the retransmission counter to "the allowable upper limit retransmission time" and subtract the "delay time accompanying retransmission" from the retransmission counter for each retransmission. By such processing, the instant when the retransmission time exceeds the allowable upper limit can be easily detected by the timing when the retransmission counter becomes negative.

Further, in the above-mentioned embodiment, the packet loss is detected by the mismatch of the order information. However, the present invention is not limited to this. For example, checksum of the header area of the packet
Packet corruption (a type of packet loss) may be detected.

In the above-described embodiment, the order information is appropriately rewritten in the relay node, so that the packet group cannot be reconstructed using this order information. Therefore, in the upper application layer or the like, it is preferable to reconfigure the packet group based on information (fragment offset or the like) different from the order information.

In addition, in the above-mentioned embodiment, it is preferable that the information newly added to the header area of the packet is added to the option portion of the conventional header area. With such consideration, it becomes possible to maintain the compatibility of the packet header configurations. As a result, even if a conventional relay node exists between the relay nodes of the present method, it is possible to perform packet retransmission between the relay nodes of the present method without any problem.

In the above-mentioned embodiment, a serial number indicating the sending order of packets is used as the order information.
However, the order information may be any information as long as it indicates the sending order of packets. For example, list-structured data designating a forward (or backward) packet may be added to the packet as order information. The packet loss can also be detected by the mismatch between the lists.
In the case of such a list structure, since it is only necessary to correct the connection relationship between the lists of the inconsistent portions, there is an advantage that the inconsistency correction can be simplified. Further, array data indicating the sending order of packets may be exchanged in advance between the relay nodes. In this case, the packet loss can be detected by detecting the mismatch between the packet actually transmitted and the array data.

Further, in the above-mentioned embodiment, when the packet loss is detected, the retransmission request is always transmitted. However, the present invention is not limited to this. For example, there may be a case where some retransmission requests are intentionally not performed due to load distribution of relay nodes. in this case,
The relay node relays directly to the subsequent stage without rewriting the "address information of the relay node" and correcting the order information. By such an operation, it becomes possible to share a part of the retransmission request in the subsequent stage.

In the above-described embodiment, step S
As described in 7, the inconsistency of the order information is corrected even when the packet is retransmitted. In this case, the packet retransmission operation does not cause apparent inconsistency in the order information. Therefore, there is an advantage that the inconsistency judgment of the order information can be easily executed.

However, in the present invention, it is not necessary to correct the inconsistency of the order information when the packet is retransmitted. In this case, the retransmission packet is inserted between the consecutive packets of the order information, which causes apparent inconsistency in the order information. Such apparent inconsistency is caused by the insertion of the retransmitted packet, and does not mean that a new packet loss has occurred. So in this case, the relay node
The order information of the retransmitted packet may be ignored as an exception, and the consistency of the order information may be determined. (The relay node itself issues the retransmission request. Therefore, the relay node
Identify the retransmission packet generated by this retransmission request,
It is possible to ignore it as an exception to the order information. In such an operation, there is an advantage that the situation in which the modification processing of the order information occurs by the domino effect due to the insertion of the retransmission packet can be prevented and the number of modification times of the order information can be reduced.

In the above description, the functions in the packet relay apparatus 11 are shared by the components shown in FIG. However, the present invention is not limited to this sharing. For example, it is possible to easily integrate the function of the resending unit 15 into the relay unit 12. That is, the present invention can be implemented in various forms without departing from the spirit or main characteristics shown in the claims.

[0062]

EFFECTS OF THE INVENTION <Claims 1 and 7> In the invention of claim 1 or 7, the relay node that temporarily stores the packet is
It adds its own address information to the packet relayed to the subsequent stage. Therefore, in the relay node, by reading the address information from the packet before or after the packet loss, the relay node that temporarily stores the lost packet can be found and the retransmission request can be accurately transmitted. With such an operation, even in a packet communication network in which the communication path is changed from time to time, packet retransmission between relay nodes can be realized with high probability. << Claims 2 and 8 >> In the invention of Claim 2 or Claim 8,
The packet loss is detected from the mismatch of the information indicating the packet sending order (hereinafter referred to as “order information”). With this detection method, it is possible to quickly detect a packet loss in the middle of packet relay without reconfiguring the packet group at the packet destination. Therefore, it is an excellent detection method for detecting packet loss in the relay node. << Claims 3 and 9 >> In the invention of Claim 3 or Claim 9,
Correct the inconsistency of the order information for the latter stage. Therefore, the packet loss is not redundantly detected in the latter stage, and it is possible to reliably prevent a situation in which retransmission requests are chained in each relay node in the latter stage. <Claims 4 and 10> In the invention of claim 4 or 10, the storage area of the relayed packet is released after the allowable delay of the packet has elapsed. By releasing the storage area after the allowable delay elapses in this way, the temporary storage of the relayed packet is guaranteed within the allowable delay, and it is possible to reliably respond to the packet retransmission. Further, it is possible to improve the situation in which the storage area is occupied forever by the relayed packets that have become unnecessary beyond the allowable delay. << Claim 5, 11 >> In the invention of claim 5 or claim 11, the retransmission counter added to the packet is updated for each retransmission. Therefore, this retransmission counter makes it possible to manage the number of retransmissions of the bucket and the retransmission time.
Further, according to the present invention, when it is determined that the number of times of retransmission or the retransmission time has reached a predetermined value or more, packet retransmission is stopped. As a result, it is possible to prevent a situation in which packet retransmission is repeated more than necessary. <Claim 6> A packet relay program according to claim 6 causes a computer to function as the relay section, the temporary storage section, the retransmission request section, and the retransmission section according to any one of claims 1 to 5. As a result, it becomes possible to realize the packet relay device according to any one of claims 1 to 5 on a computer.

[Brief description of drawings]

FIG. 1 is a diagram showing a configuration of a packet relay device 11.

FIG. 2 is a diagram showing a state of a packet communication network.

FIG. 3 is a diagram showing a data structure of a packet relayed by a packet relay device 11.

FIG. 4 is a diagram illustrating an operation of the relay unit 12.

5 is a flowchart illustrating a message handler of the resending unit 15. FIG.

6 is a flowchart illustrating a message handler of temporary storage unit 13. FIG.

7 is a flowchart illustrating a message handler of the resend request unit 14. FIG.

FIG. 8 is a diagram illustrating a packet relay operation and a packet retransmission operation.

FIG. 9 is a diagram showing an experimental network.

FIG. 10 is a graph showing a packet discard rate in this experimental network.

FIG. 11 is a diagram showing how a communication path is changed.

[Explanation of symbols]

11 Packet relay device 12 Relay section 13 Temporary storage 14 Retransmission request part 15 Retransmitter 21 Source 22 Address 23 Relay node

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Aki Kobayashi             2-12-1, Ookayama, Meguro-ku, Tokyo Tokyo Kogyo             Inside the university (72) Inventor Hideki Otsuki             3-1-12-306 Kihei-cho, Kodaira-shi, Tokyo F-term (reference) 5K030 GA12 HA08 HB12 HD03 KA04                       LA01 MB13                 5K034 AA05 AA09 EE11 FF11 HH01                       HH02 MM03

Claims (11)

[Claims] 1. A packet relay device arranged at a relay node of a packet communication network, wherein address information of its own (the relay node) is added to a packet received from a previous stage via the packet communication network. A relay unit for relaying to a subsequent stage, a temporary storage unit for temporarily storing the packet to be relayed by the relay unit, and a packet loss occurring in the packet to be relayed by the relay unit, from a packet before or after the packet loss. Read the "address information added in the previous stage",
A re-transmission request unit that transmits a re-transmission request to the preceding stage indicated by the address information, and a re-transmission unit that re-transmits the packet temporarily stored in the temporary storage unit when a re-transmission request is received from the outside. And a packet relay device. 2. The packet relay device according to claim 1, wherein the relay unit detects the packet loss from a mismatch of information indicating a transmission order of the packets (hereinafter referred to as “order information”). Packet relay device 3. The packet relay device according to claim 2, wherein the relay unit corrects the mismatch of the order information with respect to the subsequent stage. 4. The packet relay device according to any one of claims 1 to 3, wherein the temporary storage unit stores the packet after an allowable delay of the packet elapses.
A packet relay device for releasing the storage area of the packet. 5. The packet relay apparatus according to claim 1, wherein the retransmission unit updates the retransmission counter added to the packet for each retransmission, and the retransmission counter indicates the retransmission. A packet relay device, which stops retransmission when the number of times or the retransmission time exceeds a predetermined value. 6. A computer having a packet communication function is caused to function as the relay unit, the temporary storage unit, the resend request unit, and the resend unit according to any one of claims 1 to 5. Packet relay program. 7. A packet relay method implemented in a relay node of a packet communication network, comprising adding address information of itself (the relay node) to a packet received from a previous stage via the packet communication network, A relay step of relaying to a subsequent stage, a temporary storage step of temporarily storing the packet to be relayed in the relay step, and a packet loss occurring in the packet to be relayed in the relay step, starting from a packet before or after the packet loss. The address information added in the preceding stage "is read out, and a resend request step for sending a resend request to the preceding stage indicated by the address information; and a resending request received from the outside, the packet temporarily stored in the temporary storing step. And a retransmission step for retransmitting the packet. 8. The packet relay method according to claim 7, wherein in the relay step, the packet loss is detected from a mismatch of information indicating a transmission order of the packets (hereinafter referred to as “order information”). Packet relay method. 9. The packet relay method according to claim 8, wherein in the relay step, the inconsistency of the order information is corrected with respect to the subsequent stage. 10. The method according to any one of claims 7 to 9.
Item 5. The packet relay method according to item 4, wherein the temporary storage step releases the storage area of the packet after an allowable delay of the packet has elapsed. 11. The packet relay method according to claim 7, wherein the retransmission step updates a retransmission counter added to the packet for each retransmission, and the retransmission counter indicates the retransmission. A packet relay method characterized in that the retransmission is stopped when the number of times or the retransmission time exceeds a predetermined value.