JP2006050086A - Routing method and router of packet including a plurality of destination addresses - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a routing method and a router of a packet including a plurality of destination addresses in which data can be delivered while being divided or compressed appropriately. <P>SOLUTION: The routing method comprises steps of: receiving a data packet including a plurality of destination addresses through two routes of different upper limit size of data packet and then judging whether the address list of the received packet has been compressed or not; routing the data packet including a plurality of destination addresses; and outputting the data packet while dividing or compressing. The dividing/compressing step comprises steps of: judging whether the data packet subjected to routing can be transmitted as it is; dividing the data packet subjected to routing if it cannot be transmitted; and outputting the divided data packet. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a routing method and apparatus (router apparatus) in a computer network, and in particular, a routing method for a packet including a plurality of destination addresses corresponding to a difference in the upper limit value (MTU) of the packet size, and It relates to a router device.

  In a computer network, communication between terminals is realized by transmitting and receiving small pieces of data called packets to which control information such as a destination address is added. Further, in a computer network, a router device is used at a connection portion between a network and a network in order to appropriately deliver a packet from a transmission terminal to a reception terminal.

  When a packet arrives from a connected network, the router device refers to the destination address, and the destination to which the data packet is delivered based on the routing information of the control packet set in advance or sent during operation Determine the network route.

  On the other hand, a packet transfer format for realizing datagram type communication generally has a configuration having a header part including a destination address and a data part (or payload data part) for storing a main body of data to be transferred. When it is necessary to send the same message to multiple destinations, multiple destination addresses are explicitly stored in the header part within the range allowed by the capacity of one packet to reduce the consumption of network resources, and the data part is shared There is known a method of composing data packets and performing data communication.

Although the principle of operation does not change even in a router apparatus that processes a packet including a plurality of destination addresses, a packet including a single destination address is generally processed because a packet includes a plurality of destinations. In addition to typical router device routing processing (router operation), the following operations are weighted.
1) For all the destination addresses included in the destination address list of the arrived packet, a delivery destination network is determined for each address in the same manner as the routing processing of a general router device.
2) Among these destination addresses, a destination address list composed only of addresses having the same delivery destination network is constructed, and a packet including this in the header part is output to the network interface of the corresponding delivery destination.
3) The operation of 2) is repeated for each network path connected to the router device. However, the network from which the original packet arrived is excluded because the packet has already flowed on the network once.

  According to such a data communication method, the transmitting terminal can send the same data packet to a plurality of destinations, and transmits a packet that explicitly includes all destination addresses as an address list. In addition, when the receiving terminal includes its own address in the address list in the header of the packet, the receiving terminal determines that the packet is addressed to itself and performs a receiving operation.

  On the other hand, the router device may require a special operation when performing the routing process.

  In general, a network has a frame size as a unit of data transfer, that is, a specific MTU (Maximum Transmission Unit) according to the characteristics of the medium. Usually, the size of a packet to be sent over a network is determined so that it can be stored in a frame. Therefore, when the size of a packet that has arrived from a route with a large MTU is larger than the MTU of the communication route that performs the transfer, it cannot be transferred as it is.

  Therefore, in a router device that connects different network media, when the size of the arriving packet is larger than the MTU of the communication path that performs the transfer, the packet data is divided. That is, the payload data portion of the packet is divided so as to be equal to or less than the MTU of the delivery destination network, and a packet is generated by adding a header portion to which information related to the division is added to each.

The technical contents described above are described in Non-Patent Document 1, for example.
Mastering TCP / IP Introduction 3rd edition Takashi Takeshita, Koho Murayama, Toru Arai, Yukio Hirota Ohmsha (Figures 4.8, 4.12, 4.19, 4.21)

  However, in the case of a packet including a plurality of destination address lists, the size of the header portion increases in proportion to the number of destinations. Therefore, there are cases where the conventional packet data division method cannot be used.

  For example, when a packet with N destination addresses arrives from a route with a large MTU, but the MTU of the transfer destination route is small and data division is necessary, the router device first performs routing processing and performs routing processing. As a result of the processing, a header portion of a packet including a list of M destination addresses is formed in order to attempt to deliver a packet having a destination address number M (where M is a natural number equal to or less than N) for a route having a small MTU.

  At this time, there is no problem when the size of the configured header part is smaller than the MTU of the transfer destination route. However, when the size is large, the conventionally known packet data division method performs the data part division. However, the size of the divided packet never falls below the MTU. In this case, the router device cannot deliver the arrived packet and eventually has to discard the packet. As a result, there may occur a problem that data transmitted from the transmission terminal is not delivered to the destination terminal desired by the sender.

  The present invention has been made in consideration of the above-described problems. When a packet is routed between two networks having different upper limits on the size of one packet, the conventional dividing method uses the divided packet. A method for routing a packet including a plurality of destination addresses that can be delivered by appropriately dividing or compressing data even when the size of the packet is larger than the MTU of the transfer path, and the router device And

  In order to solve the above-described problem, a routing method for a packet including a plurality of destination addresses according to the present invention provides a plurality of destination addresses in at least two routes having different upper limit sizes of data packets. A multi-destination packet decompression processing procedure for executing a decompression processing step when it is compressed by determining whether an address list of the received packet is compressed after receiving a data packet including a plurality of destinations; A multi-destination routing processing procedure for performing routing processing on a data packet at an address, and a split / compression processing procedure for splitting or compressing and outputting the data packet after the routing processing as necessary. Whether the compression processing procedure can transmit the data packet after routing processing as it is A transmission permission / inhibition determining step for determining whether or not transmission is possible in the transmission permission / inhibition determining step, a division processing step for re-dividing the data packet after routing processing, and a data packet output step for outputting the data packet after division It is characterized by.

  In order to solve the above-described problem, according to the routing method for a packet including a plurality of destination addresses according to the present invention, as described in claim 2, the division processing step includes the destination address data of the header portion of the data packet. A division number determination processing step for determining the division number and the division number of the payload data portion, and a data packet division processing step for re-dividing the data packet into the division number determined in the division number determination processing step, To do.

  In order to solve the above-described problem, according to the routing method for a packet including a plurality of destination addresses according to the present invention, as described in claim 4, the division number determination processing step includes: For the two values of the number of data divisions, select the combination that minimizes the number of packets to be sent out of the two value combinations in which the size of the divided packet to be constructed is less than or equal to the upper limit of the size of one packet in the destination network It is characterized by doing.

  In order to solve the above-described problem, a router device for a packet including a plurality of destination addresses according to the present invention provides a plurality of network interfaces that receive data packets as input / output interfaces with the outside, as described in claim 6 A data packet processor for transmitting and receiving data packets, a multi-destination packet decompression processor for decompressing and outputting an address list when the address list of the data packet received by the network interface is compressed, and the network The destination of the data packet received by the interface performs a routing process for a single data packet, and the destination of the data packet received by the network interface performs a routing process for a plurality of data packets. A multi-destination routing processing unit, and a multi-destination packet splitting unit that outputs the data packet after the routing processing by dividing and compressing the data packet as necessary in consideration of the upper limit of the size of one packet of the delivery destination network. A compression processing unit, a control packet processing unit that transmits and receives control packets, and a routing table update unit that updates information of a routing table that stores routing information are provided.

  In order to solve the above-described problem, according to the router device for a packet including a plurality of destination addresses according to the present invention, the multi-destination packet division / compression processing unit is configured to perform data processing after routing processing. When a packet cannot be transmitted as it is, not only the payload / data part but also the destination address contained in the header part is tried, and multiple divided packets are created and output after the routing process. It is characterized by being configured.

  In order to solve the above-described problem, according to the router device for a packet including a plurality of destination addresses according to the present invention, as described in claim 9, the multi-destination packet division / compression processing unit performs post-routing processing. When the data packet cannot be transmitted as it is, when the data packet after routing processing is divided, the size of the divided packet to be constructed is the size of the destination network for the two values of the destination address division number and the payload data division number. It is characterized by selecting a combination that minimizes the number of packets to be sent out from a combination of two values that are equal to or smaller than the upper limit value of the size of one packet.

  According to the routing method of a packet including a plurality of destination addresses according to the present invention and the router device, when performing routing of a packet between two networks having different upper limits of the size of one packet, Even when the size of the divided packet is larger than the MTU of the transfer path, the data can be appropriately divided or compressed and delivered.

  Hereinafter, a routing method of a packet including a plurality of destination addresses according to the present invention and the router device will be described with reference to the drawings.

  Here, an embodiment of a protocol based on IPv6 as a router device and a protocol based on Xcast (eXplicity Multicast) as a transfer method for storing a plurality of destination addresses will be described. Xcast is an Internet draft proposed to IETF mainly for the purpose of efficient delivery of streaming data such as moving images (draft-ooms-xcast-basic-spec-06.txt).

  FIG. 1 shows a configuration diagram of a router device 1 that executes a routing method for a packet including a plurality of destination addresses according to the present invention.

  The router device 1 of FIG. 1 includes a plurality (n) of network interfaces 2 as input / output interfaces with the outside, and a data packet processing unit 3 that is responsible for transmission / reception of received IP data packets (hereinafter simply referred to as data packets). A single destination routing processing unit 4 that performs routing processing of a data packet having a single destination, and a multi-destination routing processing unit 5 that performs routing processing of a plurality of data packets whose destination is a plurality of data packets.

  Further, the router device 1 includes a control packet processing unit 7 that transmits and receives control packets and a routing table update unit 9 that updates information in the routing table 8 that stores routing information in order to appropriately control the routing processing. In addition.

  The network interface 2 in the router device 1 is an arbitrary network medium such as a LAN based on Ethernet (registered trademark) or a wireless network. The network interface 2 receives a packet transmitted from an external computer network. When the received packet is a data packet, the network interface 2 sends the received packet to the data packet processing unit 3.

  When receiving the data packet from the network interface 2, the data packet processing unit 3 determines whether the received packet is an Xcast packet, that is, whether the destination is single or plural. When the received packet has a single destination, the data packet processing unit 3 sends the received packet to the single destination routing processing unit 4.

  The single destination routing processing unit 4 refers to the routing table 8 stored in the router device 1 and has routing information, and sends the data packet to the reference destination data packet processing unit 3 to transmit the data packet to the delivery destination network. When the data packet processing unit 3 receives the data packet after the routing process from the single destination routing processing unit 4, the data packet processing unit 3 sends the received data packet to the network interface 2. Then, the network interface 2 delivers the data packet received from the data packet processing unit 3 to the delivery destination network.

  On the other hand, when the packet received from the network interface 2 is an Xcast packet (a plurality of destinations) in the data packet processing unit 3, the address list of the Xcast packet may be compressed. Although the specific method of compression is not specified in the Xcast specification, it is assumed that some specific compression method is adopted.

  Therefore, the router device 1 needs to be configured to be able to expand the address list of the compressed Xcast packet. Therefore, the router device 1 further includes a multi-destination packet decompression processing unit 11 that decompresses the address list of the compressed Xcast packet as shown in FIG. When there are a plurality of destinations of the packet received from the network interface 2, the packet received by the data packet processing unit 3 is sent to the multi-destination packet decompression processing unit 11.

  FIG. 2 is an example of an Xcast packet sent to the multi-destination packet decompression processing unit 11 and is an explanatory diagram for explaining the format of an IPv6 Xcast packet.

  In the present embodiment, it is assumed that a specific bit (for example, an area described as 0 in FIG. 2) of an empty area of the Xcast header can be used as means for expressing the presence or absence of compression. That is, the multi-destination packet decompression processing unit 11 can determine the presence or absence of compression by referring to the specific bit of the empty area of the Xcast header.

  Accordingly, when the multi-destination packet decompression processing unit 11 first receives an Xcast packet (data packet) from the data packet processing unit 3, the address list of the Xcast packet (data packet) is referred to by referring to a specific bit in an empty area of the Xcast header. It is determined whether or not it is compressed. As a result, if it is compressed, the address list is expanded, and the expanded Xcast packet is output to the multi-destination routing processing unit 5. If not compressed, the received Xcast packet is output to the multi-destination routing processing unit 5 as it is.

  The multi-destination routing processing unit 5 performs a routing process for distributing the Xcast packet for each network of the delivery destination by an algorithm obtained based on the Xcast specification.

  FIG. 3 is an explanatory diagram illustrating a general processing procedure for routing processing of a data packet having a plurality of destinations, and FIG. 4 is an explanatory diagram illustrating routing processing of a data packet having a plurality of destinations by a more specific example. .

  In the routing process performed by the multi-destination routing processing unit 5, as shown in FIG. 3, a destination address is assigned to each delivery destination network in the Xcast packet that was initially one. A plurality of Xcast packets having different destinations are constructed.

  Next, in the computer network as shown in FIG. 4, a router device (abbreviated as a router in FIG. 4) 1 is exemplified by a case where data is delivered from the host A to the host B, host D, host G and host H. The routing processing performed by the multiple destination routing processing unit 5 will be described.

  As a premise for explanation, it is assumed that the router device 1 shown in FIG. 4 includes a first network interface 2, a second network interface 2, and a third network interface 2 as the network interface 2. In addition, a computer network in which the first network interface 2 includes the host A, the host B, and the host C, a computer network in which the second network interface 2 includes the host D and the host E, and a third network interface 2 include It is assumed that it is connected to a computer network having a host F, a host G, and a host H.

  As a procedure for distributing a multi-destination data packet to a predetermined destination, for example, when a data packet including all destinations in the header portion is transmitted from the host A to the host B, the host D, the host G, and the host H, the transmission is performed. The received data packet is received by the terminal (host) described in the destination. First, since the host B described as the destination exists between the host A and the router device 1, the host B directly receives the data packet transmitted from the host A. Then, the router device 1 receives the data packet transmitted from the host A.

  When the router device 1 receives the data packet transmitted from the network interface 2 host A, the multi-destination routing processing unit 5 passes through the data packet processing unit 3 and the multi-destination packet decompression processing unit 11 as shown in FIG. Receives a data packet. When the multi-destination routing processing unit 5 receives the data packet, the multi-destination routing processing unit 5 starts the routing process in the procedure shown in FIG.

  In the example shown in FIG. 4, since the number of destination addresses is four, the value of N shown in FIG. 3 is N = 4. Then, with reference to the routing table 8, the distribution is performed sequentially from the first destination address to the fourth destination address to the corresponding delivery destination list (initial state is empty).

  The routing information in the routing table 8 referred to by the multi-destination routing processing unit 5 is updated by the routing table updating unit 9 that has received the control packet delivered together with the data packet, and becomes a data packet input source computer.・ Exclude the network from the delivery destination. That is, in the example shown in FIG. 4, host A, host B, and host C are excluded from the list addressed to the first network interface 2.

  Therefore, the host B as the first destination address is not included in the list addressed to any network interface 2 and is not added to the delivery destination list. Further, since the host D that is the second destination address is included in the list addressed to the second network interface 2, it is added to the delivery destination list of the second network interface 2. Further, the host G and the host H that are the third and fourth destination addresses are both included in the list addressed to the third network interface 2 and are therefore added to the delivery destination list of the third network interface 2. .

  Subsequently, it is confirmed that the delivery destination lists of the network interfaces 2 other than the first network interface 2 that is the input source network interface 2, that is, the second network interface 2 and the third network interface 2, are not empty. Then, each constructed Xcast packet is transmitted to the data packet processing unit 3 corresponding to the delivery destination list. When the multi-destination routing processing unit 5 transmits each Xcast packet to the data packet processing unit 3, the routing process of the multi-destination routing processing unit 5 is completed.

  After the multi-destination routing processing unit 5 performs routing processing, Xcast packets transmitted from the multi-destination routing processing unit 5 individually correspond to the router device 1 via the data packet processing unit 3 and the network interface 2. To the destination host D, host G, and host H. At this time, if the size of each packet exceeds the MTU of the packet delivery path, delivery cannot be performed. There is a problem.

  Therefore, as shown in FIG. 1, the router device 1 considers the size of the received data packet and the MTU of the computer network having the host (terminal) as the delivery destination, and divides the data packet as necessary. A multi-destination packet division / compression processing unit 12 that performs compression processing is further provided.

  When the multi-destination packet division / compression processing unit 12 receives a data packet having a size equal to or smaller than the MTU value of the delivery destination network, the multi-destination routing processing unit 5 first determines a data packet that does not change anything. To the data packet processing unit 3 at the delivery destination. On the other hand, when a data packet having a size larger than the MTU of the delivery destination network is received, the division or compression of the Xcast packet is weighted, and the divided or compressed data packet is output.

  The data packet processing unit 3 performs output processing on the corresponding network interface 2 as in the case of a single destination. Then, a physical packet corresponding to the network medium is transmitted from the network interface 2.

  As described above, the router device 1 includes the multi-destination packet decompression processing unit 11 so that the routing process can be performed even when the packet address list is compressed. In addition, by providing the multi-destination packet division / compression processing unit 12, when performing routing processing of packets between two networks having different upper limit sizes of one packet, the conventional division method does not perform division. Even when the packet size is larger than the MTU of the transfer path, the data can be appropriately divided or compressed for delivery.

  Next, a method for routing a packet including a plurality of destination addresses according to the present invention will be described.

  A packet routing method including a plurality of destination addresses includes a multi-destination packet decompression processing procedure for determining whether or not an address list of a received Xcast packet is compressed and executing a decompression processing step when the address list is compressed; A multi-destination routing processing procedure for performing routing processing on the data packet of the destination address, and a division / compression processing procedure for dividing and compressing and outputting the data packet after the routing processing as necessary.

  FIG. 5 is a processing flowchart for explaining in more detail the division / compression processing procedure performed by the multi-destination packet division / compression processing unit 12.

  According to FIG. 5, the division / compression processing procedure includes a division processing procedure for dividing the packet after the routing process, and a compression processing procedure for compressing the packet after the routing process. Here, the division processing procedure in the division / compression processing procedure is an indispensable processing procedure in which the packet size after the processing can be made smaller than the MTU to be delivered, while the compression processing procedure is a compression process of the address list. This is an arbitrary processing procedure that cannot guarantee that the compressed packet size is necessarily smaller than the MTU only by this compression processing procedure.

  Therefore, in FIG. 5, both the division processing procedure and the compression processing procedure are shown. For convenience of explanation, first, when the division / compression processing procedure includes only the division processing procedure which is an essential processing procedure ( The case where the compression setting parameter set in the router device 1 is “off” (corresponding to NO in step S5 described later) will be described.

  The fragmentation / compression processing procedure (division processing procedure) is a transmission permission determination step for determining whether or not the data packet after the routing process received by the multi-destination packet fragmentation / compression processor 12 can be transmitted as it is (step S1). And a division processing step (step S2 to step S3) for re-dividing the data packet when it is determined that transmission is not possible in the transmission permission / inhibition determination step (in the case of NO in step S1), and the data packet processing unit 3 A data packet output step (step S4) for outputting the data packet.

  Further, the division processing step in the division processing procedure includes the division number p (p is a natural number) of the destination address data (corresponding to a header variable length portion Hv described later) and the division number q of the payload data portion of the header portion of the data packet. A division number determination processing step (step S2) to be determined, and a data packet division processing step (step S3) for re-dividing the data packet into the division number determined in the division number determination processing step.

  The division processing procedure (step S1 to step S4) is started when the multi-destination packet division / compression processor 12 receives a packet after the routing processing (start). When the multi-destination packet division / compression processor 12 starts the division processing procedure, first, the process proceeds to step S1, and a transmission permission / inhibition determination step is performed in step S1.

  In the transmission permission / inhibition determination step, it is determined whether or not the data packet after the routing process received by the multi-destination packet division / compression processor 12 can be transmitted as it is. When it is determined that transmission is not possible in the transmission permission / inhibition determination step (NO in step S1), the process proceeds to step S2, and then a division processing step is performed. In the division processing step, first, a division number determination processing step is performed in step S2.

  FIG. 6 is an explanatory diagram outlining the division processing steps performed by the multi-destination packet division / compression processor 12.

  A data packet generally has a configuration having a header part and a payload data part. As shown in FIG. 6, this header part is a protocol-specific header required regardless of the number of destination addresses. It has a fixed length portion Hf and a header variable length portion Hv occupied by a portion that changes in proportion to the number of destination addresses.

  Accordingly, if the payload data length is D, the size of the data packet is the sum of the header fixed length portion Hf, the header variable length portion Hv, and the payload data length D (= Hf + Hv + D). In order for the multi-destination packet division / compression processor 12 to receive a data packet, the size of the received data packet (= Hf + Hv + D) is naturally smaller than the MTU of the computer network of the delivery source (hereinafter referred to as MTUin). .

  On the other hand, in order to deliver a data packet received by the multi-destination packet division / compression processor 12, the size of the received data packet must be smaller than the MTU of the delivery destination computer network (hereinafter referred to as MTUout). should not. However, the size of the data packet received by the multi-destination packet dividing / compressing unit 12 is larger than the MTU of the computer network of the delivery destination (hereinafter referred to as MTUout).

  In an IP network, IP fragmentation is known in which a payload data portion is divided to construct a plurality of packets as a specification for delivery to a computer network having an MTU smaller than the size of a received data packet. However, in the IP fragmentation, in the case of a packet that explicitly includes a plurality of destinations such as an Xcast packet, the number of bytes required for the destination address list in the packet increases in proportion to the number of destination addresses. Even if the packet is divided into a large number, it may be impossible to reduce the packet to a size smaller than the MTU of the delivery destination.

  Therefore, in order to cope with the case where data division such as IP fragmentation cannot be handled, the multi-destination packet division / compression processor 12 tries not only to divide the payload data but also to divide the address list in the header. . That is, the division number p of the header variable length portion Hv and the division number q of the payload data length D are determined (step S2), the header variable length portion Hv is divided into p pieces, and the payload data length D is divided into q pieces. p × q divided data packets are constructed (step S3).

In FIG. 6, Hv1, Hv2 to Hvp are the sizes obtained as a result of dividing the destination address into p,
[Equation 1]
Hv = Hv1 + Hv2 + ... + Hvp
It is.

On the other hand, D1, D2 to Dq are each size obtained as a result of dividing the payload data portion into q pieces,
[Equation 2]
D = D1 + D2 + ... + Dq
It is.

  Next, a method for determining the division numbers p and q of the data packet (Xcast packet) determined in the division number determination processing step of the division processing step will be described.

  In the division number determination processing step, when determining the division numbers p and q of the data packet (Xcast packet), an algorithm that provides a method for dividing one Xcast packet is adopted. As an example of an algorithm for dividing one Xcast packet (hereinafter referred to as a division number determination algorithm), first, one packet length is calculated for each combination of a destination address division number p and a data division number q. Among the combinations of p and q whose calculation result is equal to or less than MTUout, it is possible to select p and q that minimize p × q (total number of packets).

One packet length C (p, q) for each combination of p and q is
[Equation 3]
C (p, q) = Hf + B × K (N / p) + K (D / q)
Can be expressed as Here, N is the number of destination addresses, B is the number of bytes required per destination address (16 bytes for IPv6), and K (x) is a function that returns an integer value by rounding up the decimal point of x.

Note that C (p, q) shown in Equation 3 is more generally considered in consideration of the number of bytes used in padding often used in network protocols.
[Equation 4]
C (p, q) = Hf + B × K (N / p) + K (D / q) + Padding (N)
It can be expressed as. Here, Padding (N) is the number of padding bytes determined by the number of destination addresses, and is a function that can be uniquely determined according to the specifications of a correctly defined protocol.

  As an algorithm for dividing one Xcast packet, p and q that minimize p × q (total number of packets) are selected from the combinations of (p, q) in which the size of the divided data packet is equal to or smaller than MTUout. Compared to an algorithm that directly obtains the variables p and q from each parameter of Hf, Hv, D, and MTUout, the number of output packets is minimized to generate an unnecessarily large number of packets. This is to prevent it from being done.

  If the variables p and q are directly obtained from the parameters Hf, Hv, D, and MTUout, the size of the area allocated to the payload data portion also changes if the number of header divisions changes. Although it is possible to obtain p and q values that enable transfer, optimal p and q values are not always obtained from the viewpoint of minimizing the number of output packets. In other words, an unnecessarily large number of packets may be generated.

  That is, the division number determination algorithm is characterized in that an optimum combination can be calculated from the viewpoint of minimizing the number of packets to be output without depending on a change in the number of header divisions. Further, by considering Hf as the number of bytes in the header area that must be secured when the number of destination addresses is 0, it is used not only for Xcast but also for determining the number of packets that have multiple destinations. Can do.

  It should be noted that the network protocol is not limited to Xcast and generally has an optional function, and the actual value of Hf and the number of bytes B required per destination address may change depending on the presence or absence of use. Therefore, it seems that the division number determination algorithm described on the assumption that the values of Hf and B do not change cannot be applied when at least one of the values of Hf and B changes.

  However, even when the number of bytes B required per destination address changes, the specific value to be used depends on the nature of the network protocol. Always know by any of the information. Therefore, the division number determination algorithm can be used to determine the number of divisions for a packet having a plurality of destinations regardless of whether or not Hf and B change, and not limited to Xcast.

  When the multi-destination packet division / compression processing unit 12 determines the division number p of the header variable length part Hv and the division number q of the payload / data length D in the division number determination processing step, the process proceeds to step S3. A division processing step is performed. In the data packet division processing step, the multi-destination packet division / compression processing unit 12 sets p header variable length parts Hv and payload / data length D in the manner shown in FIG. 3 based on the determined values of p and q. Divide into q pieces to construct p × q pieces of divided data packets.

  Regarding the division of the payload / data portion, it is generally necessary to include information for restoring the data arrangement order in the packet. Here, it is assumed that a method according to IP fragmentation is used. In addition, for example, another area of the Xcast header can be used as information for reproducing the data arrangement order. On the other hand, with respect to the division of the destination address, no information to be reproduced is required, and since it can be delivered as an independent packet, no additional information accompanying the division is necessary.

  When the multi-destination packet division / compression processor 12 constructs p × q divided data packets in the data packet division processing step, the data packet division processing step is completed, and upon completion of the data packet division processing step, the division processing step To complete. When the division processing step is completed, the process proceeds to step S4, and a data packet output step is performed.

  In the data packet output step, all the p × q divided data packets in which the multi-destination packet dividing / compressing unit 12 is constructed are output to the data packet processing unit 3. The data packet processing unit 3 serving as an output destination of the divided data packet is the data packet processing unit 3 connected via the network interface 2 to the computer network determined by the routing processing by the multi-destination routing processing unit 5. When the data packet output step is performed, the division processing procedure ends (end).

  On the other hand, when it is determined that transmission is possible in the transmission permission / inhibition determination step (YES in step S1), the process proceeds to step S4, and the processing steps after step S4 are performed.

  If such a division processing procedure is performed, the header part (only the header variable length part Hv) is divided, and in the conventional case, the packet must be discarded, that is, the number of destination addresses is large. Even when the header length of the data packet is long and the packet cannot be reduced to a size equal to or smaller than the MTU of the delivery destination by dividing the packet only by dividing the data area, Splitting is possible.

  Next, a case where the division / compression processing procedure includes not only the division processing procedure but also a compression processing procedure will be described.

  The compression processing procedure includes a transmission permission / inhibition determination step (step S1), a compression processing step execution determination step (step S5) for determining whether or not to execute a compression processing step for performing compression processing of the address list of the data packet, A compression processing step (step S6 to step S9) for performing packet compression processing and a data packet output step (step S4) are provided. In other words, the division / compression processing procedure is a process in which a compression processing step execution determination step (step S5) and a compression processing step (step S6 to step S9) are added to the already described division processing procedure (step S1 to step S4). It is a procedure.

  Therefore, even when the fragmentation / compression processing procedure is provided with the compression processing procedure, it starts when the multi-destination packet division / compression processing unit 12 receives the packet after the routing processing, as in the case without the compression processing procedure. First, a transmission permission / inhibition determination step is performed. If it is determined that transmission is not possible in the transmission permission / inhibition determination step (NO in step S1), the process proceeds to step S5, and then a compression process step execution determination step is performed.

  In the compression processing step execution determination step, the multi-destination packet division / compression processing unit 12 determines whether or not to perform the compression processing step. Whether or not to perform the compression processing step is determined by the multi-destination packet division / compression processing unit 12 reading the value of the compression setting parameter set in the router device 1 in advance. The multi-destination packet division / compression processing unit 12 determines that the compression processing step is performed when the compression setting parameter set in the router device 1 is on (YES in step S5).

  Here, the compression setting parameter will be supplemented. The compression setting parameter set in the router device 1 can be set to off when it is predicted in advance that compression is not effective. For example, it is expected that the IPv6 address value has many common parts between addresses in the address list. In this case, a high compression rate can be expected.

  However, when a packet including an address list with few common parts is expected to arrive at a high frequency, the compression effect is small, and the processing time for trying the compression algorithm may be wasted. For this reason, as an external setting parameter of the router device 1, it is possible to select whether or not compression is performed.

  When the compression setting parameter set in the router device 1 is on (in the case of YES in step S5), the multi-destination packet division / compression processing unit 12 determines to perform the compression processing step, and proceeds to step S6. The compression processing step is performed after step S6.

  The compression processing step has a known technical content. For example, as shown in FIG. 5, a compression effect preliminary determination step (step S6) for determining the effect of data compression of the address list in advance, and data compression of the address list are performed. A data compression step (step S7) and a determination is made as to whether the size of a data packet (hereinafter referred to as a compressed data packet) constructed using the compressed address list is equal to or smaller than the MTU (MTUout) of the delivery destination computer network A compressed data packet transmission possibility determination step (step S8), and a compressed data packet construction step (step S9) for constructing a compressed data packet.

  When the multi-destination packet dividing / compression processing unit 12 completes the compression processing step as shown in FIG. 5, the process proceeds to step S4, and subsequently, the processing steps after step S4 are executed. On the other hand, when the compression setting parameter set in the router device 1 is off (NO in step S5), the process proceeds to step S2, and the multi-destination packet division / compression processing unit 12 continues to process steps subsequent to step S2. Execute.

Next, as an example of the division / compression processing procedure, assuming the following cases (1) to (4) in the router device 1 shown in FIG. 1, the division / compression processing procedure shown in FIG. Explained.
(1) The router device 1 is connected to three computer networks (a first network, a second network, and a third network), and the destination address N included in the header is 50 from the first network. The data length D of the payload data portion is 500 bytes and an Xcast packet is received.
(2) The MTU of the first network is 1500 bytes, the MTU of the second network is 500 bytes, and the MTU of the third network is 700 bytes. In some cases, it is appropriate to use “octet” as a network term, but in this description, it is all expressed in “bytes” from the viewpoint of explanation related to the configuration of the router.
(3) Of the destination addresses of the received Xcast packet, 40 destination addresses N2 to be delivered to the second network and 10 destination addresses N3 to be delivered to the third network.
(4) The compression setting parameter is set to on in the router device 1 (if the address list can be handled only by data compression, the compression processing procedure is performed and the division processing procedure is not performed).

  In the cases (1) to (4), the multi-destination routing processing unit 5 receives the Xcast packet received from the network interface 2 connected to the first network via the multi-destination packet decompression processing unit 11. To do. When the Xcast packet is received, the multi-destination routing processing unit 5 performs the routing process according to the processing procedure as shown in FIG. 3, and constructs an Xcast packet in which the destination address list is distributed for each network as the delivery destination.

The size [bytes] of one Xcast packet is generally assumed in the Xcast specification on IPv6 in the case where extended functions such as a DSCPs list and a destination extension header are not used, and Hf = 72 [bytes] and B = 16 [ Byte], the number of destinations of the Xcast packet is p, and the data is D bytes.
[Equation 5]
40 + N × 16 + 24 + 8 + D = 72 + 16 × N + D
It becomes. By using Equation 5, the size of the output packet can be calculated.

Here, the size of the Xcast packet addressed to the second network (the total size of the header and the payload data) can be calculated by substituting N = 40 and D = 500 into Equation 5.
72 + 16 × 40 + 500 = 1212, which is 1212 bytes.

Similarly, when determining the size of the Xcast packet addressed to the third network,
72 + 16 × 10 + 500 = 732, which is 732 bytes.

  That is, the multi-destination packet dividing / compression processing unit 12 receives the first packet (1212 bytes) and the second packet (732 bytes) constructed from the multi-destination routing processing unit 5, and compresses the received packets. -Perform the division procedure.

  In the compression / division processing procedure, first, a transmission permission / inhibition determination step (step S1) is performed. In the above cases (1) to (4), the multi-destination packet division / compression processor 12 receives the first packet and the second packet, and determines whether or not each received data packet can be transmitted.

  First, for the first packet, since the packet size (1212 bytes) exceeds the MTU (MTUout = 500) of the second network as the delivery destination, the multi-destination packet division / compression processor 12 It is determined that the received first packet cannot be transmitted (NO in step S1). When it is determined that transmission is impossible in the transmission / non-transmission determination step in the compression / division processing procedure (NO in step S1), the process proceeds to step S5, and then a compression processing step execution determination step is performed.

  In the compression process step execution determination step, the multi-destination packet division / compression processor 12 confirms the state (on or off) of the compression parameter. Here, since the compression parameter is set to on, it is determined in the compression processing step execution determination step that the compression processing step is executed (in the case of YES in step S5). In step S6, the multi-destination packet division / compression processor 12 performs a compression process on the first packet.

  In the compression processing step, first, a compression effect preliminary determination step is performed in step S6, and the effect of data compression of the address list is determined in advance. Whether or not there is an effect is determined by whether or not the payload data portion is larger than MTUout. This is because if the payload data portion is equal to or greater than MTUout, it is obvious that the entire size will not be smaller than MTUout no matter how much the address list is compressed. In other words, the compression effect prior determination step is a processing step for excluding data packets that cannot be delivered from data compression targets even if the data is compressed before executing the data compression.

  Here, considering the effect of compressing the address list of the first packet, the size of the header portion of the received packet is 72 + 16 × 40 = 712 bytes, and the size of the payload data portion is 500 bytes. is there. Since MTUout is 500, the multi-destination packet division / compression processing unit 12 determines that there is no effect even if the address list is subjected to data compression for the first packet (NO in step S6).

  If the multi-destination packet division / compression processor 12 determines that there is no effect even if data compression is performed on the address list (NO in step S6), the multi-destination packet division / compression processor 12 proceeds to step S2, Step S3) is performed. In the division processing step, first, a division number determination processing step is performed to determine the values of p and q (step S2). In the division number determination processing step, for example, the multi-destination packet division / compression processing unit 12 determines the values of p and q using a division number determination algorithm.

  FIG. 7 is an explanatory diagram showing the result of calculating the combination of the destination address division number p and the payload data portion division number q using the division number determination algorithm. The numerical values shown in FIG. 7 are the calculation results of C (p, q) when Hf = 72 [bytes], N = 40 [pieces], and the payload data length D = 500 [bytes].

  In the division number determination processing step, first, the multi-destination packet division / compression processor 12 calculates C (p, q) using a division number determination algorithm. Then, among the combinations of p and q that fall within MTUout (= 500), p and q that minimize the value of p × q are selected. According to FIG. 7, when p = 4 and q = 2, p × q becomes the minimum value 8, and the Xcast packet division / compression processor 12 determines p = 4 and q = 2 as the number of divisions.

  When the multi-destination packet division / compression processing unit 12 determines the values of p and q (p = 4, q = 2) using the division number determination algorithm, subsequently, a data packet division processing step is performed in step S3. The destination address is divided into four and the payload data part is divided into two. When 8 (= 2 × 4) divided packets are constructed, the data packet dividing process step is completed, and the dividing process step is completed when the data packet dividing process step is completed.

  When the division processing step is completed, the process proceeds to step S4, and a data packet output step is performed. In the data packet output step, all the eight divided data packets constructed by the multi-destination packet dividing / compressing unit 12 are sent to the data packet processing unit 3. Then, the data packet output step is completed, and the division processing procedure ends. After completing the division processing procedure, the data packet processing unit 3 that has received the divided packet sends eight divided data packets constructed by dividing the first packet to the network interface 2, and passes through the network interface 2. Deliver to the second network of the delivery destination.

  On the other hand, since the packet size (732 bytes) also exceeds the MTU (MTUout = 700) of the third network as the delivery destination for the second packet, the transmission permission determination step is performed as in the case of the first packet. Subsequent to (Step S1), a compression processing step execution determination step (Step S5) and a compression effect preliminary determination step (Step S6) are performed.

  In the case of the second packet, the size of the header portion of the packet received by the multi-destination packet division / compression processor 12 is 72 + 16 × 10 = 232 bytes, the size of the payload data portion is 500 bytes, and the total is 732 bytes. is there. In the case of the second packet, since the MTUout is 700, the second packet cannot be transmitted as it is.

  However, the area that can be allocated as a header is MTUout-D = 700-500 = 200 bytes, and it can be considered that there is a possibility of transmission by compression. Therefore, the multi-destination packet division / compression processor 12 determines that the data compression is effective for the second packet (in the case of YES in step S6). Therefore, for the second packet, the process proceeds to step S7 and a data compression step is performed.

  In the data compression step, the multi-destination packet division / compression processor 12 performs data compression of the address list (step S7). The compression rate of the address list differs depending on the packet type and the like, but here it is assumed to be 50%.

  When the data compression step is completed, a compressed data packet transmission availability determination step is subsequently performed. In the compressed data packet transmission availability determination step, the multi-destination packet division / compression processor 12 determines whether the size of the data packet constructed using the compressed address list is equal to or smaller than the MTUout of the delivery destination (step S8). .

  As a result of the compression of the address list of the second packet, the header length of the compressed Xcast packet is 72 + 16 × (50/100) × 10 = 152 bytes. As a result, the total packet length is obtained by adding 500 bytes of the payload length D to 652 bytes, which is equal to or smaller than MTUout (in the case of YES in step S8). Accordingly, the process proceeds to step S9, and subsequently, a compressed data packet construction step is performed.

  In the compressed data packet construction step, the multi-destination packet division / compression processor 12 constructs one packet including the compressed address list (step S9). Then, the multi-destination routing processing unit 5 outputs the data to the delivery destination data packet processing unit 3 (step S4). Thereafter, the compressed data packet is sent to the network interface 2 and delivered to the third network as the delivery destination via the network interface 2.

  Therefore, in the case of (1) to (4) above, the output of the data packet made from the multiple destination packet dividing / compressing unit 12 includes the compressed data packet containing the compressed address list and the divided plural packets. Divided data packets. That is, eight divided data packets obtained by dividing the first packet into eight and one compressed data packet are output from the multiple destination packet dividing / compressing unit 12.

  As described above, according to the routing method of a packet including a plurality of destination addresses and the router apparatus 1 according to the present invention, when routing a data packet to two routes having different MTUout, the data size of the received data packet is MTUout. If the data compression of the destination address information is performed, or if the data compression of the destination address information is not effective, the received data packet is divided into a plurality of packets. It can be delivered to the delivery destination.

  In particular, in the routing method for a packet including a plurality of destination addresses according to the present invention, when the division / compression processing procedure includes both the compression processing procedure and the division processing procedure, first, data compression of the destination address information is performed. If the data compression is not effective, a division processing procedure for dividing one packet into a plurality of packets can be performed.

  In the division processing procedure, by combining the division of the payload / data portion and the division of the address list (header variable length portion), conventionally, when the packet has to be discarded (the destination address list is long) Therefore, even if the size of the header part is long and packet division only by dividing the data part is not effective, the data packet can be delivered.

  In addition, in the division processing procedure, by applying a division algorithm, based on the upper limit value of the packet size in a given destination network, optimal division is performed from the viewpoint of minimizing the number of packets to be output, Packets can be delivered.

  On the other hand, in the router device 1, by configuring the compression setting parameter to be able to be switched on / off, it is possible to predict that there is no need for data compression or that a destination address group with little data compression effect can be obtained. In addition, it is possible to reduce the routing delay associated with the application of the compression algorithm.

  Note that the router device 1 does not have to include the single destination routing processing unit 4 that performs routing processing of a single data packet if the target of routing processing is always a packet including a plurality of destination addresses. . The router device 1 includes the single destination routing processing unit 4 because the received data packet may be a packet having a single destination address or a data packet including a plurality of destination addresses. Because.

The block diagram of the router apparatus which performs the routing method of the packet containing several destination addresses based on this invention. Explanatory drawing explaining the format of the Xcast packet for IPv6 which is one Example of the Xcast packet sent to the multi-destination packet expansion | extension process part of the router apparatus based on this invention. Explanatory drawing explaining the general process sequence about the routing process of the data packet which has multiple destinations. Explanatory drawing explaining the specific example in the routing process of the data packet with a several destination. FIG. 6 is a processing flowchart for explaining in more detail a division / compression processing procedure performed by a multi-destination packet division / compression processing unit of the router device according to the present invention. Explanatory drawing which outlines the division | segmentation process process which the multiple destination packet division | segmentation / compression process part of the router apparatus which concerns on this invention performs. Explanatory drawing showing the result of having calculated the combination of the division | segmentation number p of a destination address, and the division | segmentation number q of a payload data part using the division | segmentation number determination algorithm.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Router apparatus 2 Network interface 3 Data packet processing part 4 Single destination routing processing part 5 Multi-destination routing processing part 7 Control packet processing part 8 Routing table 9 Routing table update part 11 Multi-destination packet expansion processing part 12 Multi-destination packet division / Compression processing section

Claims (10)

When a data packet including a plurality of destination addresses is received through at least two paths having different upper limits of the data packet size, and then it is determined whether or not the received address list of the packet is compressed. A multi-destination packet decompression process procedure for executing a decompression process step, a multi-destination routing process procedure for performing routing processing on a data packet including a plurality of destination addresses, and dividing the data packet after the routing process as necessary A division / compression processing procedure for compressing and outputting,
This division / compression processing procedure consists of a transmission permission determination step for determining whether or not a data packet after routing processing can be transmitted as it is, and a data packet after routing processing is re-divided when transmission is not possible in the transmission permission determination step. A routing method comprising: a division processing step for performing a data packet output step for outputting a data packet after the division. The division processing step includes a division number determination processing step for determining the division number of the destination address data in the header portion of the data packet and the division number of the payload data portion;
2. The routing method according to claim 1, further comprising a data packet division processing step for re-dividing the data packet into the division number determined in the division number determination processing step.   In the division number determination processing step, when the upper limit of the packet size in the delivery destination network is exceeded, not only the division of the payload data part but also the division of the destination address included in the header part is attempted. 3. The routing method according to claim 2, wherein a plurality of divided packets are constructed and transmitted so as to be within the upper limit of the size of one packet. In the division number determination processing step, two values for the division number of the destination address and the division number of the payload data are set so that the size of the divided packet to be constructed is equal to or less than the upper limit value of the size of one packet of the delivery destination network. 3. The routing method according to claim 2, wherein a combination that minimizes the number of packets to be transmitted is selected from the combinations. When a data packet including a plurality of destination addresses is received through at least two paths having different upper limits of the data packet size, and then it is determined whether or not the received address list of the packet is compressed. A multi-destination packet decompression process procedure for executing a decompression process step, a multi-destination routing process procedure for performing routing processing on a data packet including a plurality of destination addresses, and dividing the data packet after the routing process as necessary A division / compression processing procedure for compressing and outputting,
The division / compression processing procedure includes a transmission permission determination step for determining whether or not a data packet after routing processing can be transmitted as it is,
A compression processing step execution determination step for determining whether or not to execute a compression processing step for performing compression processing of the address list of the data packet when transmission is not possible in the transmission permission determination step;
In this compression processing step execution determination step, a compression processing step that performs compression processing of the data packet when it is determined to execute the compression processing step;
A division processing step for re-dividing the data packet after the routing processing when it is determined that the compression processing step is not executed in the compression processing step execution determination step;
A routing method comprising: a data packet output step for outputting a data packet after any of the compression processing step and the division processing step is performed. Multiple network interfaces that receive data packets as input / output interfaces with the outside,
A data packet processing unit responsible for transmitting and receiving data packets;
A multi-destination packet decompression processing unit that decompresses and outputs the address list when the address list of the data packet received by the network interface is compressed;
A single destination routing processing unit that performs routing processing of a single data packet as a destination of the data packet received by the network interface;
A destination of data packets received by the network interface is a multi-destination routing processing unit that performs routing processing of a plurality of data packets;
A multi-destination packet division / compression processing unit that performs data packet division / compression processing as necessary in consideration of the upper limit of the size of one packet of the delivery destination network and outputs the data packets after the routing processing;
A control packet processor that transmits and receives control packets;
A router apparatus comprising: a routing table update unit that updates information of a routing table storing routing information. The multi-destination packet dividing / compressing unit is configured to divide a data packet after routing processing and output a divided packet when the data packet after routing processing cannot be transmitted as it is. 6. The router device described in 6. The multi-destination packet division / compression processing unit tries not only to divide the payload / data portion but also to the destination address included in the header portion when the data packet after the routing processing cannot be transmitted as it is. 7. The router device according to claim 6, wherein the router device is configured to construct and output a plurality of divided packets obtained by dividing the data packet. When the data packet after the routing process cannot be transmitted as it is, the multi-destination packet division / compression processing unit, when dividing the data packet after the routing process, has two destination address division numbers and payload data division numbers. Regarding the value, the combination of the two values in which the size of the divided packet to be constructed is equal to or less than the upper limit value of the size of one packet of the delivery destination network is selected so as to minimize the number of packets to be transmitted. The router device according to claim 8. When the data packet after routing processing cannot be transmitted as it is, the multi-destination packet division / compression processing unit first tries to compress the address list of the data packet, and still transmits the data packet even if the address list is compressed. 7. The router device according to claim 6, wherein the router device is configured to divide the data packet after the routing process when it cannot be done.

Images