WO2013035286A1

WO2013035286A1 - Group-to-group broadcast distribution path establishing method and communication devices

Info

Publication number: WO2013035286A1
Application number: PCT/JP2012/005517
Authority: WO
Inventors: 藤田　範人; 啓文植田
Original assignee: 日本電気株式会社
Priority date: 2011-09-09
Filing date: 2012-08-31
Publication date: 2013-03-14
Also published as: JP6003893B2; JPWO2013035286A1

Abstract

In a communication system, a group-to-group broadcast distribution for exchanging or sharing data among communication devices belonging to the same group is performed. Each communication device in the communication system comprises: a Bloom filter transmitting means (501) that transmits, to each of a plurality of neighboring nodes in a tree-like topology, Bloom filter values representative of a list of the identifiers of groups to which node groups existing on the side of the local node as seen from that neighboring node belong; and a group-to-group broadcast communication path establishing means (502) that establishes a forwarding table on the basis of the Bloom filter values received from each neighboring node in the tree-like topology and representative of a list of the identifiers of groups and on the basis of information of the neighboring node having transmitted the Bloom filter values.

Description

Broadcast distribution route setting method and communication device for each group

The present invention relates to a group-specific broadcast distribution route setting method and a communication apparatus for setting a broadcast distribution route for each group.

The mechanism for exchanging and sharing data between terminals (nodes) connected via a network is used by a wide range of applications. As an example of using data sharing, for example, each terminal may transmit its own location information to other terminals existing in the network. As a result, each terminal in the network can share each other's position information. There are many other forms of usage, such as exchanging and sharing sensor information (temperature, power consumption, etc.) of each terminal.

When data is exchanged and shared between terminals in a peer-to-peer manner, a method of communicating by configuring a full mesh topology (a network configuration, particularly a connection form of nodes on the network) between terminals is considered first. . However, according to this method, when there are N terminals in total, each terminal needs to transmit data to (N−1) terminals excluding its own terminal. Therefore, from the viewpoint of traffic load, when N is large, data exchange on a full mesh topology should be avoided.

Therefore, it is considered to form a logical topology between terminals and exchange data on the topology. Such a logical topology is also called a virtual topology or an overlay topology.

For example, Patent Document 1 describes an example of a method of configuring a logical topology between terminals and exchanging data on the configured logical topology. In the network communication method described in Patent Document 1, communication terminals belonging to a network are assumed to be two-level nodes of a trunk node and a branch node that communicates with another communication device through the trunk node. The logical topology is determined so that the management node can minimize the amount of communication between the core nodes. Then, each basic node communicates with other basic nodes according to the logical topology defined by the management node. Thus, bucket relay type information transmission, which is a condition for improving anonymity of the peer-to-peer network, can be performed while satisfying the improvement in communication efficiency.

When configuring a logical topology for exchanging data between terminals, generally, a redundant topology (tree topology) is obtained in order to prevent redundant transfer. Note that, in the network communication method described in Patent Document 1, the hierarchy of the tree is two hierarchies of a basic node and a branch node. On the other hand, an arbitrary number of hierarchies can be considered in the tree topology.

FIG. 11 is an explanatory diagram showing an example of broadcast delivery on a tree topology. FIG. 11 shows an example of a tree topology composed of 12 terminals (terminals T1 to T12). Here, the terminal T1 is a terminal located at the top (root) of the tree. Under the terminal T1, terminals T2 to T4 are connected as terminals belonging to a hierarchy one level below the vertex. Further, terminals T5 to T7 are connected to the terminal T2, terminals T8 and T9 are connected to the terminal T3, and terminals T10 to T12 are connected to the terminal T4. In FIG. 11, dotted arrows indicate an example of a data exchange route in the broadcast distribution on the tree topology.

In the tree topology, each terminal transmits data to all links of its own terminal when distributing the data of its own terminal to all other terminals. For example, when the terminal is the terminal T2, the terminal T2 transmits data to four terminals T1, T5 to T7.

The terminal that has received the data transfers the received data to all links other than the received link. For example, when the terminal T4 receives data from the terminal T12 in FIG. 11, the terminal T4 copies and transfers the received data to the terminals T1, T10, and T11. By repeating such an operation for each terminal, data transmitted by a certain terminal can be delivered to all terminals on the tree topology.

By the way, as a development example of a mode in which data is mutually exchanged and shared between terminals connected via a network, a method of classifying each terminal into a group and exchanging and sharing data between terminals belonging to the same group There is. For example, each terminal is grouped according to the organization to which the user belongs, and data for which each organization belongs to the distribution target is exchanged and shared only between user terminals belonging to the same organization. Patent Document 2 describes an example of a method for sharing information for each group. In the method described in Patent Document 2, clients are divided into a plurality of information sharing groups according to quality requirements. Each client manages the information sharing group, exchanges update information with other clients, and updates data for each sharing group.

Also, Patent Document 3 describes an example of a route control method for suppressing overhead during data transmission and non-transmission in an ad hoc network. The route control method described in Patent Document 3 uses a Bloom filter with the node identifier of each node existing at the link destination as a key as route information.

JP 2006-261880 A JP 2007-207013 A JP 2008-11448 A

FIG. 12 is an explanatory diagram showing an example of a method for exchanging and sharing data for each group when exchanging and sharing data for each group as described in Patent Document 2 is performed on the tree topology described above. It is. Referring to FIG. 12, each of the terminals T1 to T12 constituting the tree topology belongs to one of groups A, B, and C. In this example, terminals T4 to T6, T9, and T12 belong to group A, terminals T1, T2, T7, T10, and T11 belong to group B, and terminals T3 and T8 belong to group C.

In this example, each terminal exchanges data to be shared between terminals belonging to the same group. At that time, a route for transferring data is set for each group. That is, on a tree topology connecting all terminals, a tree-like route (broadcast delivery route) is set for each group so that data is transferred only to the terminals of the same group. In FIG. 12, three types of arrow lines indicate an example of a data exchange route (hereinafter referred to as a group-specific broadcast delivery route) in the broadcast delivery for each group.

Thus, by setting a broadcast distribution route for each group, unnecessary transfer in which only terminals belonging to different groups exist can be avoided, and traffic related to data exchange can be reduced. For example, in the example shown in FIG. 12, the broadcast distribution route of group C is only the link between the terminal T3 and the terminal T8, and it is not necessary to send traffic for data exchange to other links. Also, the group B broadcast distribution route may be only a link between terminals T1 and T2, between terminals T2 and T7, between terminals T1 and T4, between terminals T4 and T10, and between terminals T4 and T11. Therefore, this method contributes to a reduction in the amount of data traffic flowing through the network.

In order to set the broadcast distribution route for each group on the tree topology, the identifier of the group of terminals existing on the lower hierarchy terminal side from the lower hierarchy terminal to the upper hierarchy terminal in the tree topology It is necessary to notify the list of (group name, etc.). Furthermore, in this case, it is also necessary to notify the list of identifiers of the group of terminals existing on the higher hierarchy terminal side from the higher hierarchy terminal to the lower hierarchy terminal.

For example, the case of the terminal T3 in FIG. 12 will be described as an example. If the terminal T3 does not know which group belongs to each side of the terminals T1, T8, T9, the terminal T3 determines to which link the data addressed to the terminal belonging to the group C can be transferred. I can't. In order to be able to determine this, it is necessary to notify each terminal of a list of identifiers of groups of terminal groups existing on the terminal side of the upper and lower layers.

However, when the number of groups is large or the group identifier is long, another problem is that the amount of control information required for notification of the list of group identifiers for setting the broadcast distribution route for each group increases. appear. For example, when a list of group identifiers notified from one terminal to another terminal is 8 characters (8 bytes) per group and the number of groups is 100, traffic of 800 bytes is generated by one notification. Become. Terminals participating in the tree topology can change dynamically. Therefore, it is necessary to exchange lists frequently. When this notification is performed once per second, a bandwidth of 800 bytes / second (6.4 kbps) is consumed. In particular, in a situation where the available bandwidth is small, such as a wireless communication environment, such an amount of control information cannot be ignored, so it is desirable to reduce it as much as possible.

Therefore, for example, if the technique described in Patent Document 3 is used for setting a broadcast communication path for each group, the amount of control information may be reduced. However, the route control method described in Patent Document 3 is a route control method in normal unicast communication, and has the following problems when applied directly to broadcast communication for each group. That is, as described in Patent Document 3, in the method of setting all the received Bloom filter values in the routing table, the path control of the broadcast communication for each group may not function correctly depending on the network configuration. That is, the Bloom filter value is transmitted / received to / from all adjacent nodes while keeping the physical network topology as it is, and all the received Bloom filter values are set in the routing table. Then, in the case of a looping topology, Bloom filter values corresponding to the same node are received from a plurality of nodes. If all of these are set in the routing table, a packet transfer loop may occur.

For example, consider the case of a looping topology as shown in FIG. In the example shown in FIG. 13, the terminal T1 is connected to the terminal T2, the terminal T2 is further connected to the terminal T3, the terminal T3 is further connected to the terminal T4, the terminal T4 is further connected to the terminal T5, Terminal T5 is further connected to terminal T1. With this configuration, a loop topology is formed. As described in Patent Document 3, in the case of a method for performing transmission / reception processing of Bloom filter values for all adjacent nodes, for example, the Bloom filter value corresponding to the group A to which the terminal T3 belongs is changed from the terminal T3 to the terminal. The terminal T1 is notified via T2, but the terminal T1 is also notified from the terminal T3 via terminals T4 and T5. Then, the terminal T1 determines that it should transfer to both the terminal T2 and the terminal T5 as the broadcast communication path of the group A, and if this occurs every hop, it leads to a packet transfer loop.

Therefore, the present invention provides a group-by-group broadcast distribution route setting method and a communication apparatus that enable broadcast distribution for each group with a reduced amount of data traffic at the time of data distribution as well as the amount of traffic by control information at the time of route setting. The purpose is to provide. More narrowly, the purpose is to provide a group-by-group broadcast delivery route setting method and a communication device capable of setting a broadcast delivery route for each group that does not loop during data delivery with as little control information as possible. To do.

The group-specific broadcast distribution route setting method according to the present invention is a group-specific broadcast distribution route setting method for setting a group-specific broadcast distribution route for exchanging or sharing data between nodes belonging to the same group. Then, each node transmits to each adjacent node in the tree-like topology a Bloom filter value representing a list of group identifiers to which a group of nodes existing on the own node side as seen from the adjacent node belongs. Is configured to set a forwarding table based on a Bloom filter value representing a list of group identifiers received from each adjacent node in a tree-like topology and information on the adjacent node that has transmitted the Bloom filter value. .

The communication device according to the present invention is a communication device provided in a communication system that performs broadcast distribution for each group for exchanging or sharing data between communication devices belonging to the same group, and includes at least one communication via a network. Bloom filter transmission means for transmitting a Bloom filter value representing a list of identifiers of a group to which a node group existing on the own node side as viewed from the adjacent node belongs to each adjacent node in the tree-like topology connected to the apparatus And a Bloom filter value representing a list of group identifiers received from each adjacent node in a tree-like topology, and a broadcast communication for each group that sets a forwarding table based on information of the adjacent node that has transmitted the Bloom filter value Route setting means.

According to the present invention, a list of group identifiers for setting a broadcast distribution route for each group along the tree topology is exchanged using a Bloom filter. Therefore, it is possible to reduce the amount of control information necessary for notification of the list of group identifiers. Furthermore, according to the present invention, each node transfers data according to a broadcast distribution path for each group by a list expressed using a Bloom filter set along the tree topology. Therefore, the problem that the broadcast distribution data for each group loops over the network can be prevented. As a result, it is possible to reduce not only the traffic volume based on the control information at the time of route setting but also the traffic volume flowing through the network including the data traffic volume at the time of data distribution.

It is a block diagram which shows the structural example of the communication apparatus which concerns on 1st Embodiment. It is explanatory drawing which shows the example of tree structure information It is explanatory drawing which shows the relationship between group ID and a Bloom filter value. It is explanatory drawing which shows the example of a reception Bloom filter group ID storage table. It is explanatory drawing which shows the example of a forwarding table. It is a flowchart which shows the operation example of the node at the time of transmitting a Bloom filter value to an adjacent node. It is a flowchart which shows the operation example of the node at the time of receiving the Bloom filter value from an adjacent node. It is explanatory drawing which shows the example of the logical tree topology comprised from a loop-form topology. It is a block diagram which shows the outline | summary of this invention. It is a block diagram which shows the other structural example of the communication apparatus by this invention. It is explanatory drawing which shows the example of the broadcast delivery on a tree topology. It is explanatory drawing which shows the example of the broadcast delivery for every group on a tree topology. It is explanatory drawing which shows the example of a network topology.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a configuration example of a communication apparatus according to the first embodiment of the present invention. A communication apparatus 10 shown in FIG. 1 is a terminal that functions as a node in a communication system to which the present invention is applied, and includes a tree configuration unit 11, a tree configuration information holding unit 12, a Bloom filter exchange unit 13, and a forwarding table. 14, a broadcast delivery data transfer unit 15, and a Bloom filter query generation unit 16. The communication device 10 is a terminal connected to a network, and is connected to the network via a wireless interface or a wired interface. The communication device 10 is connected to other terminals via a network. Note that all terminals are assumed to have the same components as the communication device 10. Hereinafter, the communication device 10 may be simply expressed as a terminal. The link connecting the terminal to another terminal may be a logical link by encapsulation, such as an IP tunnel or a UDP / TCP tunnel, in addition to a wireless link or a wired link.

The tree configuration unit 11 configures a hierarchical tree topology between terminals. The tree configuration unit 11 exchanges control messages with other terminals to configure a tree topology applied to the terminals. Note that configuring a tree topology determines the connection form of each node on the network to which the present invention is applied so that the logical topology is a tree regardless of whether it is physically a tree. This means performing various controls for (fixing). As an example of a protocol for realizing such a function (tree topology configuration function), for example, a spanning tree protocol that is a protocol that configures a tree without a loop on Ethernet (registered trademark), or a PIM that configures a multicast tree -DM (Protocol-Independent-Multicast-Dense-Mode-). The tree configuration unit 11 registers tree configuration information indicating the configured tree topology in the tree configuration information holding unit 12.

For example, when the spanning tree protocol is used, the tree configuration unit 11 may configure the tree topology as follows. That is, a frame called BPDU (bridge protocol data と unit) is exchanged with the tree configuration unit 11 of the node on the network, and a bridge (root bridge) that is the root of the tree structure is first determined. A tree without a loop may be configured by blocking unnecessary ports based on the cost to the root bridge. More specifically, first, each bridge exchanges BPDUs and compares the bridge IDs included therein to determine the root bridge. When the root bridge is determined, each bridge further exchanges BPDUs, and calculates a path cost to the root bridge (a total value of costs determined for each link on the route). In each bridge, the port closest to the root bridge is defined as a root port (Root Port, RP). When there are a plurality of routes leading to the root bridge, the shortest route may be selected by comparing the sum of the costs obtained from the link speed. In each segment, a port closest to the root bridge (low cost) is designated as a designated port (Designated Port). Other ports are non-designated ports (Non Designated Port, NDP). Each segment may form a spanning tree by blocking non-designated ports.

Besides, it is also possible to construct a virtual network (overlay network) between nodes and virtually construct a tree topology without depending on the physical topology at all. In this case, virtual link setting may be instructed from a management server (not shown) so as to have a tree topology. Further, a logical tree topology may be configured by configuring virtual links in a distributed manner by a peer-to-peer protocol used in a distributed stream distribution tree or the like. The tree configuration unit 11 may perform such a tree configuration operation using, for example, the timing when the participation or withdrawal of another node is detected.

The tree configuration information holding unit 12 stores the tree configuration information of the tree topology configured by the tree configuration unit 11. The tree configuration information may be information indicating the connection destination (adjacent node) of the terminal together with the tree topology hierarchy information, for example. For example, the tree configuration information holding unit 12 may store a list of adjacent nodes existing on the upper side and the adjacent nodes existing on the lower side in the configured tree topology. In the case of a tree topology without a loop used for broadcast communication for each group, the tree configuration information is only information indicating the connection destination (adjacent node) of the terminal, in which hierarchical information such as upper or lower is omitted. There may be.

FIG. 2 is an explanatory diagram showing an example of an entry (tree configuration information) in the tree configuration information table registered in the tree configuration information holding unit 12. In the tree configuration information table shown in FIG. 2, tree configuration information in the terminal T2 shown in FIG. 12 is registered. This entry indicates that the terminal T1 exists as an upper side adjacent node and the terminals T5 to T7 exist as lower side adjacent nodes.

The Bloom filter exchanging unit 13 uses a Bloom filter to obtain a list of identifiers of a group to which a group of nodes existing on the own node side from each adjacent node belongs, that is, a group to which a node beyond the own node belongs, with other terminals. It is a processing unit for exchanging with each other. The Bloom filter exchange unit 13 includes a reception Bloom filter / group ID storage unit 131, a Bloom filter generation unit 132, a Bloom filter transmission unit 133, a Bloom filter reception unit 134, and a forwarding table setting unit 135.

Bloom filter is a bit string used to test whether an element is a member of a set. The Bloom filter value of an element is a value in which the key value of the element is calculated using k hash functions in an m-bit bit array, and 1 is set in the bit position corresponding to each hash value. In the present embodiment, a group identifier (group ID) to which each node belongs is expressed by a Bloom filter, and a list of group identifiers is notified with a small bit length.

FIG. 3 is an explanatory diagram showing the relationship between the group ID and the Bloom filter value in the first embodiment of the present invention. In the present embodiment, it is assumed that the Bloom filter value of group A is “01100100” as shown in FIG. Further, the Bloom filter value of the group B is “10010010”. Also, the Bloom filter value of group C is “00101001”. Note that both are 8-bit bit strings.

The reception Bloom filter / group ID storage unit 131 stores the Bloom filter value received from the adjacent node accepted by the Bloom filter reception unit 134 and the group ID of the own node.

FIG. 4 is an explanatory diagram illustrating an example of an entry (hereinafter referred to as a reception bloom filter / group ID storage table) registered in the reception bloom filter / group ID storage unit 131. FIG. 4 shows an example of a reception Bloom filter group ID storage table when the own node is the terminal T2 shown in FIG. In FIG. 4, for example, it is shown that the group ID of the own node is group B. Furthermore, the Bloom filter values received from the adjacent nodes T1, T5 to T7 are “11111111”, “01100100”, “01100100”, and “10010010”, respectively. In the present embodiment, the list of group IDs is expressed by the logical sum of the Bloom filter values of the groups included in the list.

The Bloom filter generation unit 132 generates a Bloom filter value to be transmitted to each of the other adjacent nodes based on the reception Bloom filter / group ID storage table registered in the reception Bloom filter / group ID storage unit 131. . Specifically, the Bloom filter generation unit 132 calculates the logical sum of the Bloom filter value of the group ID of the own node and all Bloom filter values received from the adjacent nodes other than the adjacent node to be transmitted. The Bloom filter value to be notified to each adjacent node is generated. The timing for generating the Bloom filter value may be set for every predetermined period, or may be the timing when the entry registered in the reception Bloom filter / group ID storage unit 131 is updated.

As an example, the own node is the terminal T2 shown in FIG. 12, and the group ID of the own node and the Bloom filter value received from each adjacent node are those shown in the received Bloom filter group ID storage table shown in FIG. Next, the case where the Bloom filter value transmitted to the terminal T1 is generated will be described. First, the Bloom filter generation unit 132 calculates the Bloom filter value of the group B that is the group ID of the own node. Here, the Bloom filter value of the group B is “10010010”. Next, the Bloom filter generation unit 132 receives Bloom filter values “01100100”, “01100100”, “01100100” received from all adjacent nodes (terminals T5, T6, T7 in this example) other than the terminal T1 that is the transmission target node. The logical sum of 10010010 "and the previously calculated Bloom filter value" 10010010 "of group B is calculated. As a result, the logical sum is “11110110”, and this value is the Bloom filter value generated as transmitted from the terminal T2 to the terminal T1.

In the above example, the own node belongs to only one group, but it may belong to a plurality of groups. For example, when the own node belongs to two groups, group B and group C, the Bloom filter value of the group ID of the own node is the Bloom filter value “10010010” of Group B and the Bloom filter value “00101001” of Group C. The logical sum is taken to be “10111011”.

Bloom filter transmission unit 133 transmits the Bloom filter value generated by Bloom filter generation unit 132 to each adjacent node. The timing of transmission may be at a preset fixed period, or may be the timing when the Bloom filter to be transmitted to the adjacent node is updated for each adjacent node.

Bloom filter reception unit 134 receives the Bloom filter value from the adjacent node on the tree topology, and registers the received Bloom filter value in reception Bloom filter group ID storage unit 131 as necessary. Also, the Bloom filter receiving unit 134 passes the identifier of the node that transmitted the Bloom filter value to the forwarding table setting unit 135 together with it.

In the present invention, when the Bloom filter receiving unit 134 receives the Bloom filter value from the adjacent node, the Bloom filter receiving unit 134 refers to the tree configuration information holding unit 12 and only receives the Bloom filter value received from the upper or lower adjacent node in the tree topology. Accept and discard bloom filter values from other nodes.

The forwarding table setting unit 135 sets a route entry in the forwarding table 14 based on the Bloom filter value passed from the Bloom filter receiving unit 134 and the identifier of the node that transmitted the Bloom filter value. When setting the route entry in the forwarding table 14, the forwarding table setting unit 135 uses the Bloom filter value passed from the Bloom filter receiving unit 134 as information indicating the group ID list Bloom filter, and the node that has transmitted the Bloom filter value The route entry having the identifier of the information indicating the adjacent node is registered. In addition, when an entry having the same adjacent node as the entry to be registered already exists, the forwarding table setting unit 135 overwrites the entry.

The forwarding table 14 is a table that is referred to in order to determine which adjacent node to transfer when the broadcast distribution data for each group is transferred. An example of entries set in the forwarding table 14 is shown in FIG. In the forwarding table shown in FIG. 5, one entry for pairing a group ID list bloom filter and an adjacent node is set. Here, the group ID list Bloom filter is a Bloom filter value representing a list of group identifiers (group ID list) to which a group of nodes existing on each adjacent node side in the tree topology belongs. By checking this Bloom filter value, it can be determined whether or not a node belonging to a certain group ID exists on the side of the corresponding adjacent node.

For example, from the example shown in FIG. 5, the group included in the group ID list represented by the Bloom filter value “11111111” on the side of the terminal T1 of the adjacent node (in this example, group A, group B, and group C). It can be seen that there are terminals belonging to. Similarly, from the example shown in FIG. 5, there is a terminal belonging to the group (group A in this example) included in the group ID list represented by the Bloom filter value “01100100” on the terminal T5 side of the adjacent node. You can see that Similarly, from the example shown in FIG. 5, there is a terminal belonging to the group (group A in this example) included in the group ID list represented by the Bloom filter value “01100100” on the terminal T6 side of the adjacent node. You can see that Similarly, from the example shown in FIG. 5, a terminal belonging to the group (group B in this example) included in the group ID list represented by the Bloom filter value “10010010” exists on the terminal T7 side of the adjacent node. You can see that

When transferring the broadcast distribution data for each group, the broadcast distribution data transfer unit 15 determines to which adjacent node the data should be transferred with reference to the forwarding table 14, and transfers based on the determination result. Broadcast distribution data for each group is transferred to the adjacent node to be transmitted. A method for determining which adjacent node should be transferred will be described. First, the broadcast distribution data transfer unit 15 generates a query for inquiry corresponding to the target group ID via the Bloom filter query generation unit 16 based on the target group ID included in the header of the broadcast distribution data. Then, the broadcast delivery data transfer unit 15 searches the forwarding table 14 for an entry in which the logical product of the group ID list bloom filter and the query for inquiry has the same value as that of the query. As a result of the search, the broadcast distribution data transfer unit 15 determines all adjacent nodes corresponding to the corresponding entry (except for the adjacent node that has transmitted the broadcast distribution data) as adjacent nodes to be transferred. . Hereinafter, an adjacent node that has transmitted the broadcast distribution data may be expressed as an “upstream adjacent node”.

The Bloom filter query generation unit 16 searches the forwarding table 14 for an adjacent node to which the data is to be transferred, from the group ID that is the broadcast distribution target of the data to be transferred by the broadcast distribution data transfer unit 15. Generate a query for The query may be a Bloom filter value created by the target group ID.

Here, consider an example in which the terminal T2 in FIG. 12 searches for an adjacent node to which the broadcast delivery message addressed to the group B received from the terminal T1 should be transferred. In this example, it is assumed that the entries in the forwarding table 14 are as shown in FIG. 5, and the Bloom filter query generation unit 16 generates “10010010” as a query corresponding to the group B. In the forwarding table shown in FIG. 5, the entries in which the adjacent nodes are terminals T1 and T7 correspond to entries in which the logical product of the query and the group ID list Bloom filter has the same value as the query. Among them, the terminal T1 is excluded because it is an adjacent node that has transmitted the broadcast distribution data to be transferred, that is, an upstream adjacent node. As a result, it is determined that the adjacent node to which the broadcast distribution data addressed to the group B from the terminal T1 is to be transferred is one of the terminals T7.

Next, the operation of the communication device (node) of this embodiment will be described. The operation for notifying other nodes of the Bloom filter value as the group ID list will be described with reference to FIG. FIG. 6 is a flowchart illustrating an operation example of a node when the communication apparatus according to the present embodiment transmits a Bloom filter value as a group ID list to an adjacent node. Referring to FIG. 6, first, the bloom filter generation unit 132 refers to the reception bloom filter / group ID storage unit 131, and the reception bloom filter / group ID table 131 stores the reception bloom filter / group ID storage unit 131. Using the contents, that is, the Bloom filter value received from another neighbor and the group ID of the own node, a Bloom filter value to be transmitted to each of the other neighboring nodes is generated (step S101). The bloom filter generation unit 132 may perform this generation operation at predetermined intervals, or each entry registered in the reception bloom filter / group ID storage unit 131 (group ID of the own group or You may perform at the timing when the reception bloom filter value of each adjacent node was updated.

As an example, the group ID of the own node and the Bloom filter value received from each adjacent node are shown in the received Bloom filter group ID storage table of FIG. 4, and the Bloom filter value transmitted from the terminal T2 of FIG. 12 to the terminal T5. The case of generating will be described. First, the Bloom filter generation unit 132 calculates the Bloom filter value of “Group B” which is the group ID of the own node. Here, it is assumed that the Bloom filter value of “Group B” is calculated as “10010010”. Next, the Bloom filter generator 132 receives Bloom filter values (“11111111”, “01100100”, “10010010”) received from all adjacent nodes (terminals T1, T6, T7) other than the terminal T5 that is the transmission target node. And the previously calculated “Group B” Bloom filter value (“10010010”). As a result, the logical sum is “11111111”, and this value is the Bloom filter value transmitted from the terminal T2 to the terminal T5.

Next, the Bloom filter transmission unit 133 transmits the Bloom filter value generated in Step S101 to each corresponding adjacent node (Step S102). In steps S101 and S102, only one adjacent node is processed. The Bloom filter transmission unit 133 may repeat the processes of steps S101 to S102 for the necessary adjacent nodes.

Next, an operation for setting a route for each node to transfer broadcast delivery data addressed to each group on the tree topology (that is, a broadcast delivery route for each group) will be described with reference to FIG. FIG. 7 is a flowchart illustrating an operation example of a node when a Bloom filter value is received from an adjacent node. Referring to FIG. 7, first, the Bloom filter receiving unit 134 receives a Bloom filter value from an adjacent node (Step S201).

Next, the Bloom filter receiving unit 134 refers to the tree configuration information holding unit 12 and checks whether or not the transmission source node of the received Bloom filter is an upper or lower adjacent node in the tree topology (step S202). , S203).

If the source node of the received Bloom filter is not an upper or lower adjacent node in the tree topology (No in step S203), the Bloom filter receiving unit 134 discards the received Bloom filter, The operation is terminated (step S205).

On the other hand, when the transmission source node of the received Bloom filter is an adjacent node on the upper side or the lower side in the tree topology (Yes in step S203), the Bloom filter reception unit 134 receives and receives the received Bloom filter. The received Bloom filter is stored in the reception Bloom filter / group ID storage unit 131 together with the information of the transmission source node (step S204).

When the Bloom filter receiving unit 134 accepts the received Bloom filter, the Bloom filter receiving unit 134 refers to the forwarding table 14, and whether an entry corresponding to the same adjacent node as the transmission source node of the accepted Bloom filter is already registered in the forwarding table 14. Whether or not is checked (steps S206 and S207).

Here, when an entry corresponding to the same adjacent node as the transmission source node of the accepted Bloom filter has already been registered in the forwarding table 14 (Yes in Step S207), the Bloom filter reception unit 134 in the forwarding table 14 The group ID list Bloom filter value of the entry is updated with the accepted Bloom filter value (step S208).

On the other hand, when the entry corresponding to the same adjacent node as the source node of the accepted Bloom filter is not registered in the forwarding table 14 (No in Step S207), the Bloom filter receiving unit 134 displays the accepted Bloom filter value. An entry having a group ID list bloom filter and having the transmission source node as an adjacent node is newly added to the forwarding table 14 (step S209).

As described above, according to the present embodiment, the list of group identifiers on the tree topology is exchanged using a Bloom filter in order to set a broadcast distribution route for each group. By using the Bloom filter, the list of group identifiers can be expressed in a small size, and the amount of control information required for notification of the list of group identifiers can be reduced.

Specifically, a group identifier can be expressed with 9.6 bits per element (group) when the false positive rate is 1%, and with 14.4 bits when the false positive rate is 0.1%. is there. For example, compared with the case where the group identifier is expressed by 8 characters (8 bytes = 64 bits), there is an effect of greatly reducing the amount of control information. In addition, when using a Bloom filter, it becomes a problem that false positive occurs. However, in the case of this embodiment, when a false positive occurs, the broadcast distribution data is also transferred to the side of an adjacent node where there is no node belonging to the target group of the broadcast distribution data (that is, misdelivery). As a result, the broadcast distribution data is finally discarded without being received as broadcast distribution data of a different group. However, since the broadcast distribution data is always transferred in the direction in which the nodes belonging to the target group exist at this time, it can be said that there is no functional problem with a slight increase in the data transfer amount.

That is, in the normal unicast route, broadcast delivery such as copying and transferring a packet to a plurality of links is not performed. Therefore, if a false positive occurs and the transfer destination cannot be determined and is transferred in the wrong direction, the packet does not reach the destination, which is a problem. However, in broadcast distribution by group, it should be avoided that broadcast distribution data is not distributed to all nodes belonging to the target group (that is, non-distribution) rather than being erroneously distributed to nodes belonging to different groups. . Therefore, the occurrence of false positives is not a fatal problem. Considering the occurrence of false positives due to the use of the Bloom filter, considering the trade-off between the increase in data traffic due to the occurrence of false delivery and the effect of reducing the amount of control messages due to the use of the Bloom filter, the appropriate Bloom The filter length should be determined.

Also, the Bloom filter is replaced to notify the group ID list. Specifically, the Bloom filter value received from other than the adjacent node on the tree topology is discarded, and only the Bloom filter received from the adjacent node on the tree topology is accepted and set in the forwarding table. In addition, this is used to calculate the Bloom filter value to send to other neighboring nodes. In this way, it is possible to set a broadcast distribution route for each group along the tree topology.

Further, according to the present embodiment, even if the topology is physically a loop topology as shown in FIG. 13, the tree configuration unit 11 logically converts a loop-free topology as shown in FIG. And the tree configuration information indicating the configuration (more specifically, the link at each node) is stored in the tree configuration information holding unit 12. By doing so, it is possible to recognize the physical topology (and physical link) and the logical topology (and logical link) separately. The Bloom filter receiving unit 134 and the Bloom filter transmitting unit 133 recognize the logical topology based on the tree configuration information stored in the tree configuration information holding unit 12. The Bloom filter receiving unit 134 and the Bloom filter transmitting unit 133 receive the Bloom filter value only from the adjacent node on the tree and do not receive the Bloom filter value from other adjacent nodes, and the Bloom filter calculated only for the adjacent node on the tree Control to send the value. Therefore, the Bloom filter value can be prevented from being transmitted and received in the loop topology as described above.

Next, the outline of the present invention will be described. FIG. 9 is a block diagram showing an outline of the present invention. A communication apparatus 500 shown in FIG. 9 is a communication apparatus included in a communication system that performs broadcast distribution for each group for exchanging or sharing data between communication apparatuses belonging to the same group, and includes a Bloom filter transmission unit 501, a group Each broadcast delivery route setting means 502 is provided. Note that the communication device 500 is connected to one or more communication devices via a network.

The Bloom filter transmission unit 501 (for example, Bloom filter transmission unit 133), for each adjacent node in the tree-like topology, lists a group identifier to which a group of nodes existing on the own node side as viewed from the adjacent node belongs. Send the Bloom filter value that represents it.

The group-specific broadcast distribution route setting means 502 (for example, the Bloom filter receiving unit 134 and the forwarding table setting unit 135) includes a Bloom filter value representing a list of group identifiers received from each adjacent node in the tree-like topology, A forwarding table is set based on the information of the adjacent node that transmitted the Bloom filter value.

By providing such a configuration for each communication device on the network, it is possible to perform broadcast distribution for each group while suppressing the amount of data traffic at the time of data distribution as well as the amount of traffic by control information at the time of route setting. .

Also, the Bloom filter transmission means 501 includes all of the Bloom filter value generated from the identifier of the group to which the own node belongs and the Bloom filter value received from each adjacent node in the tree topology other than the adjacent node to be transmitted. A logical sum with the Bloom filter value received from each adjacent node may be transmitted to each adjacent node as a Bloom filter value representing a list of group identifiers. According to such a configuration, it is possible to easily generate a Bloom filter value representing a list of group identifiers to be transmitted to each adjacent node, leading to a reduction in processing time related to notification.

Further, when the group-specific broadcast communication path setting unit 502 receives a Bloom filter value from other than an adjacent node in the tree-like topology, the received group filter value may be discarded without being set in the forwarding table. . According to such a configuration, even if a communication device that does not include the Bloom filter transmission unit according to the present invention enters the communication network, it is possible to prevent an incorrect route setting that leads to a packet transfer loop from being performed. be able to.

The communication device 500 may further include a group-by-group broadcast delivery data transfer unit 503, a tree configuration unit 504, and a tree configuration information holding unit 505. FIG. 10 is a block diagram showing another configuration example of the communication apparatus according to the present invention. The communication apparatus 500 shown in FIG. 10 further includes a group-by-group broadcast delivery data transfer unit 503, a tree configuration unit 504, and a tree configuration information holding unit 505.

The group-specific broadcast distribution data transfer means 503 (for example, the broadcast distribution data transfer unit 15), when transferring the group-specific broadcast distribution data, among the entries registered in the forwarding table, includes the bloom included in the entry. Adjacent nodes included in an entry in which the logical product of the filter value and the Bloom filter query that is the Bloom filter value generated from the identifier of the group that is the distribution target of the per-group broadcast distribution data matches the value of the Bloom filter query Then, the per-group broadcast distribution data is transferred to an adjacent node other than the upstream adjacent node of the per-group broadcast distribution data. In this configuration, the group-by-group broadcast path setting unit 502 includes a Bloom filter value that represents a list of group identifiers received from adjacent nodes in a tree-like topology, and an adjacent node that has transmitted the Bloom filter value. Correspondingly, they are registered as entries in the forwarding table.

According to such a configuration, it is possible to easily determine the transfer destination of the broadcast data for each group. Even if a false positive occurs in the Bloom filter, broadcast distribution data is always transferred in the direction in which the nodes belonging to the target group exist, so there is no functional problem with only a slight increase in the data transfer amount. .

Further, the tree configuration means 504 (for example, the tree configuration unit 11) logically configures a tree-like topology for performing broadcast distribution at least for each group.

The tree information holding unit 505 (for example, the tree configuration information holding unit 12) holds at least information indicating adjacent nodes in a tree-like topology configured by the tree forming unit. According to the configuration including the tree configuration unit 504 and the tree information holding unit 505, even if the topology is physically a loop topology, the packet caused by the fact that the group-by-group broadcast distribution route is not correctly set in the routing table. Generation of a transfer loop can be prevented.

As mentioned above, although this invention was demonstrated with reference to embodiment and an Example, this invention is not limited to the said embodiment and Example. Various changes that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention.

This application claims priority based on Japanese Patent Application 2011-197259 filed on September 9, 2011, the entire disclosure of which is incorporated herein.

The present invention can be suitably applied to a communication system that performs broadcast distribution by group for exchanging or sharing data between nodes belonging to the same group.

T1 to T12 Terminal 10 Communication device 11 Tree configuration unit 12 Tree configuration information holding unit 13 Bloom filter exchange unit 131 Reception Bloom filter / Group ID storage unit 132 Bloom filter generation unit 133 Bloom filter transmission unit 134 Bloom filter reception unit 135 Forwarding table setting Unit 14 Forwarding table 15 Broadcast distribution data transfer unit 16 Bloom filter query generation unit 501 Bloom filter transmission unit 502 Group-specific broadcast distribution route setting unit 503 Group-specific broadcast distribution data transfer unit 504 Tree configuration unit 505 Tree configuration information holding unit

Claims

A group-specific broadcast distribution route setting method for setting a group-specific broadcast distribution route for exchanging or sharing data between nodes belonging to the same group,
Each node transmits, to each adjacent node in the tree-like topology, a Bloom filter value representing a list of identifiers of groups to which a group of nodes existing on the own node side as viewed from each adjacent node belongs,
Each node sets a forwarding table based on a Bloom filter value representing a list of identifiers of the group received from each adjacent node in the tree-like topology and information on an adjacent node that has transmitted the Bloom filter value. A broadcast delivery route setting method for each group, characterized by:
Received from all adjacent nodes other than the adjacent node to be transmitted among the Bloom filter value generated from the identifier of the group to which the node belongs and the Bloom filter value received from each adjacent node in the tree topology The group-by-group broadcast delivery route setting method according to claim 1, wherein a logical sum with the Bloom filter value is transmitted to each adjacent node as a Bloom filter value representing a list of group identifiers.
Each node associates a Bloom filter value representing a list of identifiers of the group received from an adjacent node in a tree-like topology with an adjacent node that has transmitted the Bloom filter value, and registers it as an entry in the forwarding table.
Among the entries registered in the forwarding table, each node sets the forwarding destination of the broadcast distribution data to the Bloom filter value included in the entry and the identifier of the group to which the broadcast distribution data for each group is distributed Is an adjacent node included in an entry whose logical product with the Bloom filter query, which is the Bloom filter value generated from, matches the value of the Bloom filter query, and other than the upstream adjacent node of the broadcast distribution data for each group The method for setting a broadcast delivery route for each group according to claim 1 or 2, wherein the method is determined as an adjacent node.
When each node receives a Bloom filter value representing a list of group identifiers from other than adjacent nodes in a tree-like topology, the received Bloom filter value is discarded without being set in the forwarding table. Item 4. The group-by-group broadcast delivery route setting method according to any one of items 3 to 4.
Each node logically configures a tree-like topology for performing broadcast distribution at least for each group, and holds information indicating adjacent nodes in at least the configured tree-like topology,
2. Each node determines whether or not a transmission destination or a transmission source node is an adjacent node in a tree-like topology with reference to information indicating an adjacent node in the held tree-like topology. The group-specific broadcast delivery route setting method according to any one of 4.
A communication device provided in a communication system for performing broadcast distribution for each group for exchanging or sharing data between communication devices belonging to the same group,
Connected to one or more communication devices via a network,
Bloom filter transmission means for transmitting to each adjacent node in the tree-like topology a Bloom filter value representing a list of identifiers of groups to which a group of nodes existing on the own node side as seen from each adjacent node belongs;
Broadcast communication for each group that sets a forwarding table based on a Bloom filter value representing a list of identifiers of the group received from each adjacent node in the tree-like topology and information on the adjacent node that has transmitted the Bloom filter value. A communication apparatus comprising a route setting means.
Bloom filter transmission means, among the Bloom filter value generated from the identifier of the group to which the node belongs, and the Bloom filter value received from each adjacent node in the tree-like topology, all adjacent nodes other than the adjacent node to be transmitted The communication apparatus according to claim 6, wherein a logical sum with the Bloom filter value received from is transmitted to each adjacent node as a Bloom filter value representing a list of group identifiers.
A group-specific broadcast path setting means associates a Bloom filter value representing a list of identifiers of the group received from an adjacent node in a tree-like topology with an adjacent node that has transmitted the Bloom filter value, and a forwarding table As an entry for
When transferring broadcast distribution data for each group, among the entries registered in the forwarding table, the Bloom filter value included in the entry and the identifier of the group to which the broadcast distribution data for each group is distributed Adjacent nodes included in an entry whose logical product with the Bloom filter query that is the generated Bloom filter value matches the value of the Bloom filter query, and adjacent to other than the upstream adjacent node of the per-group broadcast distribution data The communication apparatus according to claim 6 or 7, further comprising a group-specific broadcast distribution data transfer unit configured to transfer the group-specific broadcast distribution data to the node.
The broadcast transmission path setting unit for each group discards the received Bloom filter value without setting it in the forwarding table when receiving the Bloom filter value from other than an adjacent node in the tree-like topology. The communication device according to any one of 8.
Tree constructing means for logically constructing a tree-like topology for performing at least broadcast distribution for each group;
The communication apparatus according to claim 6, further comprising: a tree information holding unit that holds information indicating an adjacent node in at least a tree-like topology configured by the tree configuration unit.