JPH0964913A - Data guarateeing method for packet communication - Google Patents

Abstract

PROBLEM TO BE SOLVED: To realize a data guaranteeing method for packet communication in which data are guaranteed more quickly without causing a useless traffic and delay. SOLUTION: In the case of sending a transmission packet 130 while being divided into one fragment 150 or over, a decoding fragment 170 with the data length of each fragment 150 and guaranteeing the loss of each fragment 150 is generated corresponding to the fragments 150 and the fragment 170 is sent to a transmission destination similarly to each fragment 150. When a loss takes place in any fragment 160 at a receiver side, a restoration fragment 180 is used to re-configure the received packet 140.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a packet communication data guarantee method, and more particularly to a packet communication data guarantee method for guaranteeing a packet lost in a packet network.

[0002]

2. Description of the Related Art Generally, when various types of data are communicated between information processing apparatuses via a packet network, the contents of the data communication are determined to be correct communication such as file transfer, and When a loss occurs, the data is guaranteed by retransmitting a packet corresponding to the lost data. However, in such a data guarantee method, the receiving side detects the data loss, notifies the transmitting side, and resends the lost data from the transmitting side in response to this notification, which increases the time required for communication. However, it is not suitable as a data guarantee method for communication that requires real-time processing.

Conventionally, as a data guarantee method for solving such a problem, when a packet is lost for a predetermined number of packets, data for restoring these packets is transmitted in advance as another packet. If a packet loss actually occurs, a data guarantee method has been proposed in which the packet loss is restored on the receiving side based on the data previously received in another packet for restoration. (Eg "Gigabit Networkin
g Research at Bellcore ", IEEE Network March 1992).

[0004]

Therefore, the conventional data guarantee method for packet communication has the following problems. First of all, since a data guarantee packet is created and transmitted for a large number of packets, if a packet loss actually occurs, data guarantee is not performed until the data guarantee packet arrives. Therefore, there is a problem that it takes time to guarantee the data. Secondly, since the amount of data to be transmitted at one time differs depending on the type of communication network, when data is guaranteed based on the number of packets or a fixed data length, useless traffic and delay occur. was there.

Thirdly, since data is guaranteed based on the number of packets, it is difficult to guarantee data adapted to the data loss rate of each communication network when passing through communication networks having different data loss rates. There was a problem. The present invention is intended to solve such a problem, and an object of the present invention is to provide a data guarantee method of packet communication that can guarantee data more quickly without causing unnecessary traffic and delay. There is.

[0006]

In order to achieve such an object, the data guarantee method for packet communication according to the present invention is a guarantee for guaranteeing the loss of each fragment created from a packet on the transmitting side. If you create a decompression fragment with data and send it like each fragment, and the receiving side receives each fragment and the decompressing fragment from the sending side, and one of the fragments is lost In this method, a lost fragment is restored by using the restoration fragment, and the original packet is reconstructed from the restored fragment and other received fragments.

Therefore, on the receiving side, if any of the fragments from the transmitting side is lost, the lost fragment is restored using the restoring fragment sent from the transmitting side in the same manner as each fragment, The original packet is reconstructed from this restored fragment and other received fragments.

Further, on the transmission side, the number of restoration fragments for a packet to be transmitted is controlled based on the data loss rate of the packet communication network used for packet communication. Therefore, on the transmitting side, the restoration fragments are created and transmitted for the packet to be transmitted by the number based on the data loss rate of the packet communication network.

[0009]

Next, the present invention will be described with reference to the drawings. FIG. 1 is an explanatory diagram showing a data guarantee method for packet communication according to an embodiment of the present invention.
FIG. 4 is a block diagram showing a data communication system using the packet network data guarantee method of the present invention, and FIG. In FIG. 1A, reference numeral 100 denotes a CPU 120 that performs arithmetic processing, a memory 121, and a communication network interface (hereinafter referred to as communication network I / F) 122.
Is an information processing device including a communication network I / F 122, is connected to the communication network 110, and is connected to another information processing device 100.
Can communicate with.

FIG. 1B shows the configuration of the memory 121 of each information processing apparatus 100, and the transmission packet 13
Three fragments 150 and 160 are provided for 0 and the received packet 140, respectively, and restoration fragments 170 and 180 are provided. Generally, when sending and receiving data, the data length that can be sent and received is specified according to the communication network and communication protocol used, and the send and receive data that has that data length is called a fragment, and the send and receive packets are In the case of a long packet, the transmitted / received packet is divided into a plurality of fragments, stored, and transmitted / received.

In this case, the transmission packet 130 is divided and stored in three fragments 150, and the communication network I /
It is transmitted to the communication network 110 via F122. Further, the contents of the three fragments 160 received from the communication network 110 via the communication network I / F 122 are combined to generate the reception packet 140. Restoration fragment 170
Is a fragment generated by a predetermined logical operation from the three fragments 150, and the fragment 15
It is transmitted / received in the same manner as 0 and 160, and is used as restored data when any of the fragments 160 is lost.

These packets and fragments have a structure as shown in FIG. 2, for example. Note that FIG.
Then, it is a diagram showing an example of a packet format including information for distinguishing a packet, a fragment, and a fragment for restoration, and does not limit the packet format. As a premise, 1 packet
When it is possible to transmit in one fragment, the packet is to be transmitted in one fragment.

In FIG. 2, an identification flag 210 indicates a value for identifying a fragment or a fragment for restoration,
As a result, switching between normal fragment processing and restoration fragment processing, which will be described later, is performed. The fragment end flag 211 indicates whether or not the fragment is the last fragment, and is used for determining whether or not all the fragments forming the packet have been received.

The packet length 212 indicates the length of the fragment in the case of a fragment, and sets the total length of the packet to be restored in the case of a fragment for restoration to set the length of the last fragment constituting the packet. It becomes possible to judge the loss. Packet identifier 2
Reference numeral 13 indicates an original packet, a fragment, or a restoration fragment.

When the identification flag 210 indicates a fragment, the offset 214 indicates from which position in the original packet the fragment is, and when the identification flag 210 indicates a restoration fragment, a plurality of restoration flags are used. It shows an identifier for a fragment, and can be used even when there are a plurality of restoration fragments. The transmission source address 215 indicates the communication network address of the information processing apparatus that transmits the packet, and the transmission destination address 216 indicates the communication network address of the information processing apparatus to which the packet should be transmitted. Data 22
0 indicates communication data.

Next, the operation of the present invention will be described with reference to FIGS. FIG. 3 is a flowchart showing a processing operation at the time of packet transmission, and FIG. 4 is a flowchart showing a processing operation at the time of packet reception. When sending a packet,
As shown in FIG. 3, as one of packet fragmentation, one transmission packet 130 is divided into one or more fragments and stored (step 310).

After the fragmentation is completed, the transmission packet 13
From each fragment 150 that composes 0, a restoration fragment 170 is created based on a predetermined fragment creation method described later (step 320). Then, after transmitting each fragment 150 to the communication network 110 via the communication network I / F 122 (step 330), the fragment for restoration 170 is similarly transmitted (step 340),
The transmission of one transmission packet 130 is completed (step 350).

Here, a method of creating a fragment for restoration will be described. Note that the method described below is an example of a method for creating a restoration fragment, and other methods may be used to create a restoration fragment. Here, as shown in FIG. 5, consider the case where the packet 500 is divided into three fragments 510, 520, and 530, and one restored fragment 540 is created.

First, the packet header 505 is copied into the fragment headers 511, 521 and 531 of the three fragments 510, 520 and 530 and the header 541 of the restored fragment 540. Then, the data portions 501, 502, and 503 of the packet 500 are converted into the fragment 51.
0, 520, 530 are divided and stored so as to have a predetermined size. The final fragments 530 do not have to have the same size.

Next, each fragment header 511, 52
1,531, the contents of the data portion 501,50
2, 503 according to the data stored. Furthermore, the data portion 501, 502, 5 of each fragment
The exclusive OR for each bit of 03 is restored data 54
Let it be 2. At this time, when the sizes of the data portion 503 of the last fragment 530 do not match, the bits of the insufficient portion are regarded as 0, and the exclusive OR is taken to obtain the restored data 542.

Therefore, when the fragments 510, 520, 530 and the decompression fragment 540 created by this method are transmitted and the loss of any one of the fragments occurs, the data of all the other fragments received and the data for decompression are recovered. It becomes possible to derive the data of the lost fragment by taking the exclusive OR of the restored data of the fragment.

On the other hand, when receiving a packet, as shown in FIG. 4, first, an arbitrary fragment is received from the communication network 122 via the communication network I / F 122 (step 410), and the identification flag 210 of the fragment is referred to. The type of the received fragment is determined (step 411). Here, when the received fragment is other than the fragment for restoration (step 411: NO), it is determined whether the packet can be reassembled from the already received fragment 160 by referring to the fragment end flag 211. (Step 412).

As a result, when the fragments 160 are insufficient (step 412: NO), the process returns to step 410 and waits for the reception of the fragments 160. If packet reconfiguration is possible (step 412: YE
S) Then, the received packet 140 is reconstructed based on these fragments 160 (step 420), and the series of packet reception processing ends.

On the other hand, in step 411, when the received fragment is the restoration fragment 180 (step 411: YES), the corresponding packet is already re-established by referring to the packet identifier 213 of the received restoration fragment 180. It is judged whether or not it has been configured (step 431), and if it has already been reconfigured (step 431: YES), this fragment for restoration 180 is discarded (step 432), and a series of processing upon packet reception is completed. To do.

In step 431, if the packet has not been reassembled yet (step 431: N
O), the lost fragment 160 is derived using the restoration fragment 180 to reconstruct the received packet 140 (step 433). After this, the received packet 14
It is determined whether or not the packet reconfiguration is successful based on the error correction information of 0 (step 434), and if the packet reconfiguration is not successful (step 434: NO), after setting the packet reception error (step 434). In step 450), similarly to the case where it succeeds (step 434: YES), the series of processes upon packet reception is ended.

As described above, when the transmission packet is divided into one or more fragments and transmitted, a fragment for restoration having the data length of the fragment and guaranteeing the loss of the fragment is created corresponding to these fragments. In addition to sending each fragment to the destination as well as each fragment, if any fragment is lost at the destination, the recovery packet is used to reconstruct the received packet. It is possible to guarantee, and a faster data guarantee is realized without causing unnecessary traffic and delay.

Next, referring to FIG. 6, as a second embodiment of the present invention, a case will be described in which data is guaranteed in accordance with the data loss rate of the communication network used. FIG.
It is a flowchart which shows the processing operation at the time of packet transmission adapted to the data loss rate of the communication network used.

First, the number of fragments and the number of fragments for restoration are determined for the packet to be transmitted (step 601). For example, the data loss rate of the communication network is 1
When represented by the average amount of data that is lost when data is lost once, the number of data loss L that can occur when a packet of a predetermined size is transmitted is the number of data loss L = packet size / data loss rate. Becomes

Here, when one restoration fragment is generated for a plurality of fragments, the lost fragment can be restored from the remaining fragments and the restoration fragments only when any one of them is lost. Therefore, by setting the number B of restoration fragments required per packet equal to the number of fragments desired to be restored per packet, that is, the data loss number L, it becomes possible to transmit and receive packets without loss.

Therefore, the number of restoration fragments B per packet adapted to the data loss rate of the communication network is as follows: number of restoration fragments B = number of data loss L = packet size / data loss rate. Further, the number of fragments F required to transmit one packet is: number of fragments F = packet size / fragment size.

Therefore, the number of fragments used to generate one restoration fragment, that is, the number of target fragments N to be restored by one restoration fragment, is: target fragment number N = (fragment number F) / ( The number of fragments for restoration B) = (packet size / fragment size) / (packet size / data loss rate) = data loss rate / fragment size. Actually, these calculation results are rounded up to an integer and used.

On the basis of the fragment number F calculated in this way, a fragment is created by dividing the transmission packet into each fragment and storing them (step 602), and the fragment number B for restoration is also calculated. And a fragment for restoration is created from these fragments based on the target fragment number N (step 603) and transmitted to the communication network (step 6).
04). The method of creating the fragment and the fragment for restoration is the same as described above.

As an actual calculation example, for example, the number of fragments F required for one packet is 6 from the relationship between the packet size and the fragment size, and the number of data loss times L from the relationship between the packet size and the data loss rate. Is 1
In this case, the number of restoration fragments B = the number of data loss L = 1, the number of target fragments N = F / B = 6, and one restoration fragment may be generated and transmitted for every six fragments. In this case, one decompression fragment is transmitted per packet.

As another calculation example, for example, the number of fragments F required for one packet is 6 from the relationship between the packet size and the fragment size, and the number of data loss is calculated from the relationship between the packet size and the data loss rate. L
Is 2, the number of restoration fragments B = the number of data loss L = 2, the number of target fragments N = F / B = 3, and one restoration fragment may be generated and transmitted for every three fragments. In this case, two restoration fragments are transmitted per packet.

As described above, the number B of restoration fragments and the number N of target fragments for guaranteeing packets are based on the data loss rate of the communication network used during packet communication.
Is calculated, and the fragment for restoration is created and transmitted based on these numerical values, so that data guarantee adapted to the data loss rate of the communication network used can be performed.

[0036]

As described above, according to the present invention, the transmitting side creates the restoration fragment having the guarantee data for guaranteeing the loss of each fragment created from the packet, and the same as each fragment. And then the receiving side receives each fragment and restoration fragment from the transmission side, and if any of each fragment is lost, the restoration fragment is used to restore the lost fragment. , Since the original packet was reconstructed from this restored fragment and other received fragments, compared with the conventional method of creating and transmitting packets for data guarantee for a large number of packets It becomes possible to guarantee data in units of packets, without causing unnecessary traffic and delay. Rapid data assurance is achieved.

Further, since the number of restoration fragments for a packet to be transmitted is controlled on the transmitting side based on the data loss rate of the packet communication network used for packet communication, the number of conventional packets is used as a reference. As compared with the method of guaranteeing data, it is possible to realize the guarantee of data adapted to the data loss rate of each packet communication network.

[Brief description of drawings]

FIG. 1 is a block diagram showing a data communication system using a data guarantee method for a packet network according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing a configuration example of a packet and a fragment.

FIG. 3 is a flowchart showing a processing operation during packet transmission.

FIG. 4 is a flowchart showing a processing operation at the time of receiving a packet.

FIG. 5 is an explanatory diagram showing a method of creating a fragment and a fragment for restoration.

FIG. 6 is a flowchart showing a processing operation at the time of packet transmission adapted to a data loss rate of a communication network as another embodiment of the present invention.

[Explanation of symbols]

100 ... Information processing device, 110 ... Communication network, 120 ... CP
U, 121 ... Memory, 122 ... Communication network interface (communication network I / F), 130 ... Transmission packet, 140 ... Reception packet, 150, 160 ... Fragment, 170,
180 ... A fragment for restoration.

Claims (2)

[Claims] 1. A transmitting side divides a packet into one or more fragments and transmits the same, and a receiving side reconstructs an original packet from these fragments received via a packet communication network, thereby transmitting the packet. In packet communication in which packets are exchanged between the sender and the receiver, the sender creates a restoration fragment having guarantee data for guaranteeing the loss of each fragment created from the packet, and the same as each fragment. The receiving side receives each fragment and the decompressing fragment from the transmitting side, and if any of the fragments is lost, the decompressing fragment is used to reconstruct the lost fragment, It will reassemble the original packet from this decompressed fragment and other received fragments. Data guarantee method of packet communication, characterized in that the 2. The data guarantee method for packet communication according to claim 1, wherein the transmitting side controls the number of restoration fragments for the packet to be transmitted based on the data loss rate of the packet communication network used for packet communication. A data guarantee method for packet communication characterized in that