JP2005065100A - Data distribution method, repeating system, and computer program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To achieve improvement of distribution quality of each receiver by performing data distribution with redundancy required for each receiver, respectively in consideration of heterogeniety of a communication path, when packet loss generated on the communication path is restored by applying FEC to multicast. <P>SOLUTION: Each receiver 40 calculates the necessary optimal redundancy on the basis of measured packet loss rate and requests it to the upstream. In each repeating device 20, the redundancy required by each device on the downstream is stored from the received redundancy requests, the redundancy requests are integrated and transferred to the upstream as one redundancy request. A transmitter 10 performs multicast distribution of data with the redundancy according to the received redundancy request. In each repeating system, the data from the upstream is adjusted to the redundancy required by each device on the downstream and transferred. Thus, the data distribution with the redundancy requested by every receiver, respectively is made possible. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a data distribution method for repairing a packet loss occurring in a communication path and improving distribution quality in multicast in which the same data is distributed from a transmission apparatus to a plurality of reception apparatuses.

  In recent years, there is an increasing demand for applications that require multicast, such as multi-point video streaming. As a protocol for realizing multicast in an IP (Internet Protocol) network based on packet communication, IP multicast, application layer multicast, and the like (see, for example, Patent Document 1) have been proposed. In these systems, a packet distributed from a sender is delivered to a plurality of receivers by duplicating in a multicast router arranged in the network.

  However, in a best effort type network such as the Internet, delivery packets are frequently lost, which may lead to degradation of delivery quality. In order to improve this quality degradation in multicast, several schemes have already been proposed (see Non-Patent Document 1, for example). Many of these methods are based on ARQ (Automatic Repeat Request) which is repaired by retransmitting a packet lost in a communication path. In ARQ, when a packet loss occurs, a retransmission request is transmitted from the receiver to the retransmitter, and the retransmitter that has received this retransmits the lost packet to the receiver. In other words, at least a round-trip transmission delay between the receiver and the retransmitter is necessary for retransmission, and a data reception delay becomes large in a large-scale network such as the Internet. In addition, when the number of recipients increases, the frequency of retransmission requests increases, and the processing load on the retransmission parties increases.

  On the other hand, there is a method called FEC (Forward Error Correction) that repairs a lost packet by using an error correction code. In FEC, data to be transmitted is encoded and transmitted, and even if a packet loss occurs in the communication path, the original data is restored by decoding at the receiver. In FEC, lost packets can be repaired only by processing within the receiver, so that reception delay can be reduced compared to ARQ described above. Further, since retransmission is not required, the load is not concentrated on a specific node, and the scalability with respect to the number of recipients is excellent. However, on the other hand, since the amount of data increases due to encoding, the distribution bandwidth also increases, and congestion may occur depending on the state of the communication path, resulting in an increase in packet loss. Furthermore, when there are a large number of lost packets, decoding cannot be performed, and lost packets may not be repaired.

As an FEC encoding method, RS (Reed-Solomon) code, LT (Lubic Transform) code, or the like is used. As an example, this RS code has the properties of a systematic code and an optimal code.
The systematic code is a property that the original data is included in the encoded data. That is, when k packets are encoded with the RS code, the original k packets and l redundant packets are obtained. Here, the ratio (l / k) between the number of redundant packets and the original number of packets is called redundancy. That is, the greater the redundancy, the greater the number of redundant packets. An RS code can be encoded by setting an arbitrary redundancy.
The optimum code is a property that a minimum number of redundant packets is required to repair a certain amount of lost packets. Specifically, when l redundant packets are generated, up to l lost packets can be repaired. When one or more packets are lost, decoding cannot be performed and the lost packets cannot be repaired.

As described above, if the redundancy is too small, the lost packets cannot be repaired. If the redundancy is too large, the distribution bandwidth increases and the number of lost packets also increases due to congestion. As a result, there is an optimum value for the redundancy that varies depending on the communication path conditions such as the packet loss rate and the communication path capacity. In order to effectively use the FEC, this redundancy setting value is important.
When one-to-one communication is considered, in a public network such as the Internet, the channel condition changes with time due to fluctuations in background traffic, and the optimum redundancy also changes with time. Therefore, a method called AFEC (see Non-Patent Document 2) that adaptively changes the redundancy has been proposed. In AFEC, an optimum redundancy is obtained based on a packet loss rate periodically measured at the receiver, and the redundancy is used for encoding processing at the sender to cope with fluctuations in communication channel conditions. The algorithm for finding the optimal redundancy is basically enough redundancy to repair the packet loss that has occurred, but if the packet loss increases even if the redundancy is increased, it is excessive. It is judged that congestion has occurred due to the redundancy, and the redundancy is reduced. Thereby, the optimum redundancy that minimizes the packet loss rate is adaptively set.

  On the other hand, considering the case where FEC is applied to multicast that is one-to-many communication, the setting of redundancy becomes even more difficult. In an inhomogeneous network such as the Internet, communication path conditions differ for each recipient, and the optimum redundancy also differs. Here, a problem arises if the same redundancy is distributed to all recipients. For example, if the redundancy is increased in accordance with a recipient having a large packet loss rate, the distribution band increases, and a recipient having a small communication path capacity increases the packet loss due to congestion. On the other hand, if the redundancy is reduced, a recipient with a large packet loss rate will not have enough redundancy, and repair of the packet loss will be impossible. Even if the average packet loss rate is adjusted to a receiver, a receiver with insufficient redundancy and a receiver that causes congestion appear.

Further, in Patent Document 2, paying attention to the difference in packet loss rate between the wired link and the wireless link, the distribution data is encoded by FEC in the relay apparatus installed immediately before the wireless link, and the FEC is applied only to the wireless link. A method has been proposed that absorbs the difference in communication channel status between a wired link and a wireless link. However, the load of the FEC encoding process is large, and when the number of flows to be encoded increases, the load on the relay apparatus also increases.
JP 2003-032300 A JP 2003-174478 A JC Lin and S. Paul, “RMTP: A reliable multicast transport protocol (RMTP)”, Proc. IEEE INFOCOM 1996 (IEEE Infocom 1996) , P. 1414-1424, March 1996 K. Park and W. Wang, “AFEC: An adaptive forward error correction protocol for end-to-end transport of real-time traffic (AFEC) In Proc. Int. Conf. Computer Communications and Networks, Lafayette, LA (Lafayette, Louisiana, USA), October 1997

As described above, when FEC is applied to multicast, there is a problem that a receiver whose packet loss rate is increased appears if distribution is performed with the same redundancy to each receiver due to a difference in communication channel conditions for each receiver. is there.
The present invention has been made in consideration of this problem, and even when the channel condition is inhomogeneous, the delivery quality can be improved for each receiver, and the low repair delay and high scalability, which are the advantages of FEC, can be achieved. An object of the present invention is to provide a data distribution method, a relay device, and a computer program that can improve distribution quality in multicast.

In order to solve the above-mentioned problem, the invention according to claim 1 is used in a communication network that performs multicast distribution in which the same data is distributed from a transmitting device to a plurality of receiving devices via one or more relay devices, and is redundant. A data distribution method for repairing packet loss that occurs in a communication path by an encoding method, wherein the transmission device distributes data with redundancy according to a downstream relay device or reception device. It is.
According to the present invention, in the case of multicast distribution in which the same data is distributed from a transmitting device to a plurality of receiving devices via one or more relay devices, when packet loss that occurs in a communication path using FEC is repaired, By changing the redundancy in each relay device and distributing the data with different redundancy for each receiver, it is possible to absorb the difference in the communication path status for each receiver.

The invention according to claim 2 is the data distribution method according to claim 1, wherein the relay device receives the redundancy requested by each downstream relay device or receiving device and stores it internally. The upstream transmission device or relay device is notified of the redundancy obtained by aggregating a plurality of received redundancy.
According to the present invention, the relay device receives a plurality of redundancy requests sent from a plurality of downstream relay devices and receiving devices, stores the redundancy required by each device, and stores a plurality of redundancy requests. And a single redundancy request is transmitted to an upstream device, so that load concentration of redundancy request processing in the transmission device can be prevented.

The invention according to claim 3 is the data distribution method according to claim 2, wherein the relay device has the highest redundancy among the redundancy required by each of the received downstream relay devices or receiving devices. The redundancy is collected by selecting.
According to the present invention, the redundancy request aggregation processing performed by the relay device selects the largest redundancy requested by each of the plurality of downstream relay devices and the receiving device, and transmits it to the upstream device. Thus, it is possible to request data with sufficient redundancy for all downstream apparatuses.

The invention according to claim 4 is the data distribution method according to any one of claims 1 to 3, wherein the distribution data is encoded according to the redundancy received from the downstream relay device, and the downstream It distributes to the relay device or receiving device.
According to the present invention, since the transmission device distributes data based on the redundancy received from the relay device, it is possible to request data with sufficient redundancy from all downstream devices.

The invention according to claim 5 is the data distribution method according to any one of claims 1 to 4, wherein the relay device is configured to provide redundancy of data received from the transmitting device or an upstream relay device. The degree of redundancy is adjusted to each redundancy requested by a downstream relay device or receiving device, and the data is transferred.
According to the present invention, the relay device adjusts the redundancy of the data distributed from the transmission device according to the redundancy requested from the plurality of downstream relay devices or reception devices, and the redundancy that differs for each downstream device. By transferring the data with, it is possible to absorb the difference in the communication channel status for each receiver.

The invention according to claim 6 is the data distribution method according to any one of claims 1 to 5, wherein the relay device is configured to distribute data received from the transmission device or an upstream relay device. The redundancy is adjusted by transferring a different number of packets to be used for each downstream relay device or receiving device.
According to the present invention, the redundancy adjustment process in the relay apparatus is performed by transferring a different number of packets distributed from the sender for each downstream apparatus, and does not require an encoding process and has a small load. The feature is that the redundancy can be adjusted.

  The invention according to claim 7 is the data distribution method according to any one of claims 1 to 6, wherein the redundancy is determined by a redundancy request packet having information on the redundancy in a payload. An upstream device is notified.

  The invention of claim 8 is a relay apparatus used for multicast distribution in which the same data is distributed from a transmitting apparatus to a plurality of receiving apparatuses via one or more relay apparatuses, and the upstream transmitting apparatus or A relay apparatus comprising: a redundancy adjustment unit that distributes data distributed from a relay apparatus using a redundancy encoding method with redundancy according to a downstream relay apparatus or reception apparatus.

  The invention according to claim 9 is the relay apparatus according to claim 8, wherein the redundancy requested by each of the relay apparatuses or downstream receiving apparatuses is received and stored therein, and the plurality of received redundancy It further includes a redundancy request processing unit that notifies the upstream transmission device or relay device of the redundancy obtained by integrating the degrees.

  The invention of claim 10 is a computer program executed on a relay device used for multicast distribution in which the same data is distributed from a transmitting device to a plurality of receiving devices via one or more relay devices. And causing the computer to execute the step of distributing the data distributed from the upstream transmission device or relay device using the redundancy coding method with the redundancy according to the downstream relay device or reception device. It is a computer program.

  The invention of claim 11 records a computer program executed on a relay device used for multicast distribution in which the same data is distributed to a plurality of receiving devices from one transmitting device via one or more relay devices. A computer-readable recording medium that distributes data distributed from an upstream transmission device or relay device using a redundancy encoding method with redundancy according to a downstream relay device or reception device; A recording medium for recording a computer program to be executed by a computer.

According to the present invention, when repairing a packet loss that occurs on a communication path by applying FEC to multicast, when transmission is performed with the same redundancy to all receivers due to the heterogeneity of the communication path status Has the effect that distribution quality can be improved for each receiving device by distributing data with the required redundancy for each receiving device in response to the problem of receivers whose distribution quality deteriorates. .
In addition, compared with the conventional ARQ-based distribution quality improvement method, using FEC as a base has the effect of realizing low repair delay and high scalability.
Compared with the method of adjusting redundancy by FEC encoding in the relay device as in Patent Document 2, the present invention adjusts redundancy only by controlling the number of transfer packets in the relay device. There is an effect that the processing load is low and the scalability for the number of flows handled at the same time is high.

Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 shows a distribution form of a communication system according to an embodiment of the present invention. Here, packet communication in the IP network is assumed. Then, data distributed from the transmission device 10 is received by the plurality of reception devices 40 via the plurality of relay devices 20 and the multicast router 30 using the multicast protocol. The multicast protocol is the aforementioned IP multicast or application layer multicast. The multicast router 30 is a router corresponding to the multicast protocol to be used. The relay device 20 is a router corresponding to a multicast protocol used for distribution, and a device having an additional function described later.

FIG. 2 is a block diagram showing functional expansion of the inside of each device constituting the communication system according to the embodiment shown in FIG.
The multicast router 30 is an existing router having a multicast protocol, and is omitted because it is not a newly proposed device. However, as in FIG. 1, between the transmission device 10 and the relay device 20, the relay device 20 It may be installed between the receiving devices 40.

The transmission device 10 includes, for example, a data generation unit 11 that reads video and audio data from an HDD (Hard Disk Drive), a camera, and the like to generate distribution data, and a packet that generates a multicast packet by dividing the distribution data into appropriate sizes. The generation unit 12, the FEC encoding unit 13 that encodes the generated packet by FEC (Forward Error Correction), and the communication that distributes the encoded packet downstream and receives the redundancy request from the downstream device The interface (I / F) 14 includes a redundancy control unit 15 that stores the received redundancy request and instructs the FEC encoding unit 13 to perform encoding according to the redundancy request.
In FEC, data to be distributed is encoded for every k packets to generate l redundant packets, and a total of n (= k + 1) packets are transmitted. If any k packets out of n packets can be received on the receiving side, the original k packets can be restored by decoding. Here, (l / k) is the redundancy.

The relay device 20 receives a packet distributed from an upstream device, and also duplicates the communication I / F 21 that transmits an aggregated redundancy request packet upstream and the number of downstream devices to be transferred. The packet duplicating unit 22, the redundancy adjusting unit 23 for adjusting the redundancy of the duplicated packet, and transmitting the redundancy adjusted packet to the downstream device and receiving the redundancy request packet from the downstream device Stores the redundancy request from the communication I / F 24 and each received downstream device, instructs the redundancy adjustment unit 23 to perform redundancy adjustment according to the redundancy request, and aggregates a plurality of redundancy requests The redundancy request processing unit 25 generates a redundancy request packet having the aggregated redundancy in the payload.
Note that the payload indicates a portion in which an actual data body to be originally transferred is set, excluding a header portion indicating address information of a destination device among packets transmitted and received between the devices. That is, the packet is composed of “header” + “payload”.

  The receiving device 40 receives a packet distributed from an upstream device, transmits a generated redundancy request packet to the upstream device, and an FEC decoding unit that decodes the received packet by FEC 42, a packet receiving unit 43 that extracts distribution data from the decrypted packet, a data receiving unit 44 that outputs the extracted distribution data to a display device if it is video data, for example, and packet loss from the received distribution packet. A packet loss detection unit 45 that detects the packet loss rate and obtains a required redundancy from the measured packet loss rate, and generates a redundancy request packet having the redundancy in the payload A portion 46 is provided.

  FIG. 3 shows an example of a distribution form according to the present invention, and an example will be described. As shown in the figure, a case is considered where data is distributed from the transmission device 10 to three reception devices 40 (these are referred to as reception devices 40a, 40b, and 40c). A packet for data distribution is multicast-distributed from the transmission device 10 to the reception device 40a via the relay device 20, and to the reception devices 40b and 40c via the relay device 20 and the multicast router 30 downstream thereof. The

  First, multicast transmission of distribution data from the transmission device 10 to the relay device 20 is started. Here, for example, an RS (Reed-Solomon) code is used for the encoding process in the FEC encoding unit 13 of the transmission apparatus 10. Also, since the redundancy of the encoding process has not yet received a request for redundancy from the downstream, an appropriate value determined in advance, such as 1%, is selected or encoding is not performed.

The relay device 20 duplicates the packet distributed from the sender's transmitter 10 and forwards it to the receiver's receiver 40 a and the multicast router 30. Here, in the redundancy adjustment processing in the redundancy adjustment unit 23, since the redundancy request from the downstream has not yet been received, the redundancy adjustment is adjusted to an appropriate redundancy determined in advance, such as 1%, or the redundancy adjustment. Do not do.
The multicast router 30 duplicates the distribution data transferred from the relay device 20 and transfers it to the receiving devices 40b and 40c.

  The receiving devices 40a, 40b, and 40c receive the packet transferred from each upstream device, decode it, obtain the original distribution data, and output it. Further, the packet loss is detected from the received packet. The packet loss is detected, for example, by monitoring the sequence number in the received packet by the receiving device 40 by including a sequence number when generating a packet in the transmitting device 10 and checking whether the number is interrupted. This can be achieved. For example, the packet loss rate is obtained from the average value of the number of lost packets periodically every predetermined time, such as every second, and further, the redundancy requested to the upstream is obtained based on the packet loss rate, and the redundancy is requested. Generate a packet and send it to an upstream device. For example, the same method as the AFEC described above is used as a method for determining the required redundancy from the packet loss rate.

Here, for example, 5% packet loss occurs in the multicast router 30 or in the link between the relay device 20 and the multicast router 30 due to the influence of background traffic and noise, and between the relay device 20 and the reception device 40a. Consider a case in which almost no packet loss occurs in the link, but the link capacity is limited to a small amount due to the influence of background traffic. In this case, the receiving device 40a requires a low redundancy, for example 1%, while the receiving devices 40b and 40c require a large redundancy, for example 7%.
These redundancy request packets are received by the relay device 20, and the redundancy requested by the receiving devices 40a, 40b, and 40c is stored. Then, the redundancy requests are collected and transferred to the transmission device 10. Since the redundancy aggregation process selects the maximum required redundancy, a redundancy request of 7% reaches the transmission device 10.

In accordance with the redundancy requirement, the transmission apparatus 10 changes the redundancy of encoding by FEC to 7% and distributes it.
Here, when forwarding to all the receiving devices 40 with a redundancy of 7%, the link capacity between the relay device 20 and the receiving device 40a is limited to a small size, so that congestion occurs due to excessive redundancy, and the receiving device The packet loss rate of 40a may become large. This phenomenon becomes more prominent as the distribution scale increases.
Therefore, in the present invention, in the relay device 20, the distribution data is adjusted and transferred to the redundancy required for each downstream device. That is, the redundancy is reduced to 1% and transferred to the receiving device 40a, and the packet is transferred to the receiving devices 40b and 40c with the same redundancy of 7%. Here, the process of reducing the redundancy is performed by limiting the transfer amount of the redundant packet. Specifically, when transferring to the receiving device 40a, it is performed by transferring only (l / 7) of the received redundant packets. As a result, it is delivered to each receiving apparatus with the required redundancy.
Here, even if the packet loss rate fluctuates due to the influence of background traffic fluctuations, etc., the redundancy is improved so as to increase the lost packet repair rate at each receiver by periodically updating the redundancy request in each receiver 40. Is changed.

  Note that, in a configuration in which a plurality of relay devices 20 exist between the transmission device 10 and each reception device 40, a device one hop downstream of the relay device 20 can be the relay device 20 instead of the reception device 40. To work.

  Note that the transmission device 10, the relay device 20, the multicast router 30, and the reception device 40 described above have a computer system therein. The above-described operation process is stored in a computer-readable recording medium in the form of a program, and the computer system reads and executes this program to perform the above processing. The computer system here includes an OS and hardware such as peripheral devices.

  In addition to ROM, “computer-readable recording medium” refers to a portable medium such as a magnetic disk, a magneto-optical disk, a CD-ROM, a DVD-ROM, or a storage device such as a hard disk built in a computer system. That means. Furthermore, the “computer-readable recording medium” refers to a volatile memory (RAM) inside a computer system that becomes a client or a system when a program is transmitted via a network such as the Internet or a communication line such as a telephone line. In addition, those holding programs for a certain period of time are also included.

The program may be transmitted from a computer system storing the program in a storage device or the like to another computer system via a transmission medium or by a transmission wave in the transmission medium. Here, the “transmission medium” for transmitting the program refers to a medium having a function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication line) such as a telephone line.
The program may be for realizing a part of the functions described above. Furthermore, what can implement | achieve the function mentioned above in combination with the program already recorded on the computer system, and what is called a difference file (difference program) may be sufficient.

It is a figure which shows the delivery form of the communication system by one embodiment of this invention. It is a block diagram of each apparatus which comprises the communication system by the embodiment. It is a figure which shows an example of the delivery form when the packet loss by the same embodiment and the link capacity restriction | limiting have generate | occur | produced.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 ... Transmitting device 11 ... Data generation part 12 ... Packet generation part 13 ... FEC encoding part 14, 21, 24, 41 ... Communication I / F
DESCRIPTION OF SYMBOLS 15 ... Redundancy control part 20 ... Relay apparatus 22 ... Packet duplication part 23 ... Redundancy adjustment part 25 ... Redundancy request processing part 30 ... Multicast router 40, 40a, 40b, 40c ... Receiving apparatus 42 ... FEC decoding part 43 ... Packet receiver 44 ... Data receiver 45 ... Packet loss detector 46 ... Redundancy request generator

Claims (11)

Data distribution that is used in a communication network that performs multicast distribution in which the same data is distributed from a transmitting device to a plurality of receiving devices via one or more relay devices, and that repairs packet loss that occurs in a communication path using a redundant encoding method. In the method
The transmitting device is
Deliver data with redundancy according to downstream relay device or receiving device,
A data distribution method characterized by the above. The relay device is
Receive and store the required redundancy from the downstream relay device or receiving device,
Notifying the upstream transmission device or relay device of the redundancy obtained by aggregating a plurality of received redundancy levels,
The data distribution method according to claim 1. The relay device selects the largest redundancy among the redundancy required by each received downstream relay device or reception device, and aggregates the redundancy.
The data distribution method according to claim 2. The transmission device encodes distribution data according to the redundancy received from the downstream relay device, and distributes the data to the downstream relay device or reception device.
The data distribution method according to any one of claims 1 to 3, wherein the data distribution method is performed. The relay device adjusts the redundancy of the data received from the transmission device or the upstream relay device to the respective redundancy requested from the downstream relay device or the reception device, and transfers the data.
The data distribution method according to any one of claims 1 to 4, wherein the data distribution method is performed. The relay device adjusts redundancy by transferring a different number of packets used for data distribution received from the transmitting device or upstream relay device for each downstream relay device or receiving device.
The data distribution method according to any one of claims 1 to 5, wherein The redundancy is notified to an upstream device by a redundancy request packet having the redundancy information in the payload.
The data distribution method according to any one of claims 1 to 6, wherein the data distribution method is performed. A relay device used for multicast distribution in which the same data is distributed from a transmitting device to a plurality of receiving devices via one or more relay devices,
A redundancy adjustment unit that distributes data distributed from an upstream transmission device or relay device using a redundancy encoding method with redundancy according to a downstream relay device or reception device;
A relay device comprising: Redundancy request processing for receiving the redundancy requested by each downstream relay device or receiving device, storing the received redundancy internally, and notifying the upstream transmitting device or relay device of the aggregated redundancy Part,
The relay apparatus according to claim 8, further comprising: A computer program executed on a relay device used for multicast distribution that distributes the same data to a plurality of receiving devices via one or more relay devices from a transmitting device,
A step of distributing data distributed from an upstream transmission device or relay device using a redundancy encoding method with redundancy according to a downstream relay device or reception device;
A computer program for causing a computer to execute. A computer-readable recording medium recording a computer program executed on a relay device used for multicast distribution that distributes the same data to a plurality of receiving devices from one transmitting device via one or more relay devices. And
A step of distributing data distributed from an upstream transmission device or relay device using a redundancy encoding method with redundancy according to a downstream relay device or reception device;
A recording medium for recording a computer program for causing a computer to execute.

Images