JP2011145749A - Peer information providing system for p2p type communication - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To follow network information update inside/outside an AS, concerning a P4P service where a P2P distribution control server and an in-AS network information server perform efficient distribution in cooperation with each other. <P>SOLUTION: A request peer being a user which receives candidate peer information of desired content from the distribution control server transmits a peer list request to a peer information providing system 1. The peer information providing system 1 periodically receives route information to a prefix to the inside/outside of the AS from a router, manages the prefix in the latest state by association with a segment ID based on the route information, generates a peer list by making selection from the candidate peers based on the request peer and the segment IDs to which the candidate peers belong, and then provides the list to the user. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a system that provides information on a peer as a communication partner to a peer that participates in a P2P network and performs P2P type communication by using network information managed by the peer, and varies particularly on a daily basis. The present invention relates to a system that can provide peer information by following route information.

  In recent years, content delivery using P2P (Peer to Peer) technology has increased traffic between ASs (Autonomous Systems) and inside ISPs (Internet Services Providers), which has severely pressured the backbone. . In response to this technical issue, P4P (Provider Portal for P2P), where multiple ASs disclose network information and use this information on the P2P delivery service side, and AS and the delivery service cooperate with each other for efficient delivery. ) Is undertaken. P4P is disclosed in Non-Patent Document 1.

  In P4P, iTracker (network information server) installed inside AS provides peer information with good distribution efficiency for its network to appTracker (distribution control server). When a peer (requesting peer) requests content, appTracker notifies the iTracker of information on the peer (candidate peer) that it owns as peer information that has a piece of the requested content and can be connected as a P2P communication target. iTracker calculates the network distance between the requesting peer and the candidate peer based on its own network information, preferentially selects the one with the short distance, and returns it to appTracker as a peer list. Here, as the network information, a PID indicating a set of address spaces predetermined by the operator or an AS number is used. Request peers and candidate peers are controlled so that communication occurs on the same network to some extent locally, with the ratio of 70% being within the same PID and 80% being within the same AS. Delivery is realized.

P4P: Provider Portal for Applications http://www.cs.washington.edu/homes/arvind/papers/p4p.pdf

  In the above prior art, it is assumed that a prefix which is PID and address space information belonging to each PID is statically given. However, there are thousands to tens of thousands of prefixes managed within an AS, and it is considered that prefixes are added, deleted, and aggregated on a daily basis. It is not easy for a person to manually update PID information.

  In the prior art, it is assumed that a prefix belonging to an external AS is also given statically. For example, for the correspondence between the prefix and the AS number, information published on the Internet can be used. However, since this information is provided on a file basis, there is a problem that it is not possible to follow a path change that has occurred until it is updated.

  The object of the present invention is to solve the above-mentioned problems of the prior art, and P2P based on PID and prefix automatically managed and updated following a vast amount of network information that fluctuates on a daily basis within the local AS and external AS. It is to provide a peer information providing system for P2P type communication that can provide peer information of communication.

  In order to achieve the above-described object, the present invention provides a request peer that performs P2P (peer-to-peer) type communication in a network and requests a desired content, transmits the ID of the desired content to the distribution control server, and transmits each ID of the desired content. Receiving candidate peer information having piece pieces, receiving a peer list request in which the requesting peer has transmitted the candidate peer information together with address information of the requesting peer, and from the candidate peer information based on the peer list request The peer information providing system for peer-to-peer communication that provides the requesting peer with a peer list that selects peers to be communicated is characterized by the following measures.

  (1) Receiving means for receiving route information including a route to a prefix to which the requesting peer belongs and a route to a prefix to which a candidate peer in the peer list belongs from a router operating in the network. And a classifying unit that classifies a prefix included in the route information by using the updated route information and associates it with a segment ID each time the route information is received and updated by the receiving unit.

  (2) The classification unit classifies the prefix based on network identification information to which the prefix belongs, included in the route information, using the route information, and associates the prefix with a segment ID.

  (3) Further, the segment ID of each peer is determined as the segment ID associated with the prefix to which each peer belongs, and the peer to be communicated is selected from the candidate peer information based on the segment ID of each peer Selecting means for creating the peer list.

  (4) The selection unit preferentially selects candidate peers having a matching segment ID of the requesting peer and a candidate peer segment ID included in the candidate peer information, and creates a peer list.

  (5) The selection means includes priority setting means for setting a predetermined priority for a set of the segment ID of the requesting peer and the segment ID of the candidate peer, and the candidate peers are arranged in descending order of priority. To create the peer list.

  (6) The router includes an OSPF router, and the receiving means receives the route information from the OSPF router via the LSA monitoring node of the OSPF area included in the network, and the route information includes the network of the OSPF area. If the summary LSA is included and the prefix belongs to the OSPF area, the classification means associates the segment ID with each OSPF area.

  (7) When the router includes a BGP router, the route information includes a BGP UPDATE issued by the BGP router, and the classification unit includes, for each Origin AS, the prefix belonging to the Origin AS in the BGP UPDATE. Associate segment IDs.

  According to the present invention, the following effects are achieved.

  (1) Peer list request by a user (requesting peer) who obtains candidate peer information of desired content from the distribution control server in P4P, where AS and distribution service (distribution control server) cooperate with each other to efficiently distribute both The peer information providing system 1 receiving the network information can receive the route information of the network from the router, and can manage the prefix associated with the segment ID in the latest state. In addition, the peer information providing system 1 does not require manual setting by the operator when updating a large amount of network information including its own AS and external AS.

  (2) Since the prefix is associated with the segment ID based on the belonging AS or AS configuration in the route information, the prefix reflecting the network state can be managed based on the AS information.

  (3) As mentioned above, the prefix associated with the segment ID is managed in the latest state, and the peer information providing system 1 determines the segment ID of each peer as the segment ID of the prefix to which each peer belongs. In addition, since the peer list is created based on the segment ID of each peer, the peer list can provide the user with a good distribution efficiency reflecting the latest state of the network.

  (4) Since candidate peers whose segment IDs match the request peer are preferentially selected in the peer list, it is possible to provide a peer list with high distribution efficiency in which the request peer and candidate peer belong to the same prefix.

  (5) Since a predetermined priority is set for each pair of segment IDs of a requesting peer and a candidate peer and a peer list is created based on the priority, distribution is performed with priority between predetermined segments (or within segments). Can provide a peer list.

  (6) Since a segment ID for each OSPF area is generated, a peer list based on the OSPF area to which the peer belongs can be provided.

  (7) Since the external AS obtained from the BGP UPDATE is recognized by the Origin AS and a segment ID for each external AS is generated, a peer list based on the external AS to which the peer belongs can be provided.

It is a figure which shows the structure of the peer information provision system in this invention. It is a figure which shows the example of a connection structure in the case of using route reflection within AS. It is a figure which shows the example of a connection structure when a BGP router is connected by a full mesh within AS. It is a figure which shows the example of a connection structure in case the some BGP router is connected using confederation within AS. It is a figure which shows the structural example in the case of collecting OSPF route information within AS. It is a figure which shows the example of network group information. It is a figure which shows the procedure which produces / updates network group information. It is a figure which shows the example of the predetermined priority defined with respect to the group of request | requirement peer SID and candidate peer SID.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 shows the configuration of a peer information providing system according to the present invention. The peer information providing system 1 includes a route receiving unit 2, a network information management unit 3, and a server unit 4. The route receiving unit 2 collects route information from a router 5 operating in a network composed of a single or a plurality of ASs for which the P2P distribution service is implemented.

  Note that the network including the router 5 includes a plurality of P2P communication peers (requesting peers and simultaneously candidate peers) indicated by user # 1, user # 2, etc. in FIG. A content server (not shown in the figure, which may be a user who plays an equivalent role instead of a content server in Pure P2P) that accumulates content that is a piece piece of content such as audio, and the piece piece for each piece of content A distribution control server (not shown in the figure) that manages the address information of the peers that hold the P2P distribution service is possible by this network and the peer information providing system 1.

  The network information management unit 3 updates the network group information using the route information received from the router 5 at any time by the route receiving unit 2. The server unit 4 sends a request for a peer list listing candidate peers holding pieces of pieces of desired content from each user (request peer) such as user # 1, user # 2, etc. The IP address of the request peer and candidate peer information described later (It is not information that considers network efficiency in piece piece distribution). The server unit 4 collates the candidate peer information and the requested peer IP address with the network group information of the network information management unit 3, and preferentially selects candidate peers having a network distance close to the requesting peer (predetermined rate). As a result of the selection, the list is returned to the requesting peer as the user as peer list information (information in consideration of network efficiency in piece piece distribution). The user refers to the peer address for each piece of content in the peer list information, connects to the peer, downloads the piece, and obtains the desired content.

  Candidate peer information that the user transmits to the server unit 4 together with the peer list request includes a piece of requested content from the distribution control server by the user transmitting the requested content ID to the distribution control server in advance. It is received as information on existing peers. The candidate peer information is information that is periodically updated and managed by the distribution control server, such as the content held by the peer, the piece piece held, the address, and the connection possibility, and includes the correspondence information between the peer and the piece piece. However, it is not information that considers network efficiency in piece distribution to the requesting peer.

  Note that, instead of the requesting peer receiving the candidate peer information and transmitting it to the peer information providing system 1 after the requesting peer has requested the candidate peer information holding the desired content from the distribution control server as described above, the distribution control is performed. The server may send the candidate peer information directly to the peer information providing system 1 without returning it to the requesting peer. In this case, the distribution control server also receives the address information of the requesting peer from the requesting peer in advance and transmits it to the peer information providing system 1. Further, in this case, it can be considered that one of the users such as user # 1 and user # 2 corresponds to the distribution control server in FIG. That is, the distribution control server transmits a peer list request as a user to the peer information providing system. In this case, the peer list returned by the peer information providing system 1 may or may not be routed through the distribution control server in the middle of reaching the requesting peer.

  In addition, when a request peer first downloads content for which a piece of content does not exist in a peer group participating in the P2P distribution service, all content will be downloaded from the content server (or content providing user). After that, the piece of the content is distributed to the network peers, and the address information of the peers is registered and managed in the distribution control server. In addition, every time there is a request for downloading the content from various arrangement request peers on the network, the piece of the content is widely distributed to the peers on the network and is registered and managed in the distribution control server. The content for which the peer information providing system 1 according to the embodiment of the present invention provides a peer list is such that pieces of pieces are distributed and distributed to P2P distribution service participating peers, and the addresses of the participating peers and owned content IDs are distributed to the distribution control server. It is assumed that (identifier) and piece piece ID of the content are registered and managed.

  The connection between the router 5 and the path receiving unit 2 and the connection between each user # 1, user # 2, etc. and the server unit 4 may not be the same data link. In particular, in FIG. 1, each user # 1, user # 2, etc. and the server unit 4 are drawn separately from the network including the router 5, but the peer list request and transmission / reception of the peer list information with the server unit 4 are made clear. For the sake of convenience, as described above, each user # 1, user # 2, etc. is actually arranged in a network composed of a single or plural ASs ahead of the router 5. Further, each user # 1, user # 2, etc. does not have to belong to the same AS as the peer information providing system 1 in particular.

  Next, each part of the peer information providing system 1 will be described in more detail.

  First, the route receiving unit 2 receives route information from the router 5 according to the AS routing protocol and network configuration in which the router 5 operates (1) using BGP (Border Gateway Protocol), and (2) OSPF. The case where (Open Shortest Path First) is used and the case where (1) and (2) are used together (3) will be broadly described.

  In the route reception method (1), the router 5 is a BGP router, and the peer information providing system 1 is connected to this BGP router by IBGP (Internal Border Gateway Protocol), and BGP UPDATE propagated in the AS as IBGP as route information. Collect (route updates). In this method (1), BGP information is collected. Therefore, unlike the method (2) described later, not only the route within the local AS (the router 5 to which the peer information providing system 1 is connected) but also the external AS All routes can also be collected. However, in the method (1), since the route within the self AS is a unit route aggregated for the external AS, route information divided into network groups as finely as the method (2) cannot be obtained.

  In addition, (1) When using BGP, the connection configuration of the BGP router in its own AS, (1-1) When using route reflection in the AS, (1-2) The BGP router in the AS is fully meshed. (1-3) the case where a plurality of BGP routers are connected using confederation within the AS. As will be described below, in any of these cases, it is possible to collect route information by collecting BGP UPDATE.

  (1-1) An example of a connection configuration when using route reflection in the AS is shown in FIG. The router 5 to which the peer information providing system 1 is connected is a BGP route reflector 110 and is connected by IBGP. The route reflector 110 is connected to the border routers 111, 112, and 113 by IBGP. The border routers 111, 112, and 113 are connected to an external AS of the local AS 11 by EBGP (External Border Gateway Protocol). The border routers 111, 112, 113 receive the route information from the external AS of the own AS 11 and the route information inside the own AS 11 by BGP UPDATE, and transmit these to the route reflector 110. The peer information providing system 1 collects all routes of the internal AS and the external AS of the local AS 11 by collecting BGP UPDATE of the route reflector 110.

  (1-2) FIG. 3 shows an example of a connection configuration when a BGP router is connected with a full mesh in the AS. The router 5 to which the peer information providing system 1 is connected is all the BGP routers 121, 122, and 123 in the own AS 12, and they are connected by IBGP. All the BGP routers 121, 122, and 123 in the own AS 12 are connected to each other by IBGP and EBGP connected to the external AS of the own AS 12 respectively, and the route information of the own AS 12 and the external AS is received by BGP UPDATE. The peer information providing system 1 collects BGP UPDATEs of all the BGP routers 121, 122, and 123 in the own AS 12 to collect all routes inside and outside the own AS 12.

  (1-3) FIG. 4 shows an example of a connection configuration when a plurality of BGP routers are connected using confederation within the AS. The own AS 13 to which the peer information providing system 1 is connected is divided into a plurality of sub-ASs 1310, 1320, 1330 by BGP confederation, and the router 5 to which the peer information providing system 1 is connected is divided into a plurality of sub-ASs 1310, 1320, It belongs to any one of 1330. Although FIG. 4 shows a case where the arbitrary one is a sub-AS 1310 and the sub-AS 1310 is configured by route reflection, the sub-AS 1310 may be configured by a full mesh.

  When the sub-AS 1310 has a route reflection configuration, the router 5 to which the peer information providing system 1 is connected is the route reflector 131 in the sub-AS 1310 and is connected by IBGP as in the case of FIG. The route reflector 131 is connected to the border routers 1311, 1312, and 1313 through IBGP. The border router 1311 makes an EBGP connection with an external AS of its own AS 13, and the border routers 1312 and 1313 make an EBGP connection with the sub-ASs 1320 and 1330, respectively. The sub-ASs 1320 and 1330 are also EBGP-connected, and the border routers 132 and 133 are EBGP-connected to the external AS. Through these IBGP connection and EBGP connection, the peer information providing system 1 collects the routes of the local AS 13 and the external AS by collecting BGP UPDATE of the route reflector 131.

  When the sub-AS 1310 has a full mesh configuration, the router 5 to which the peer information providing system 1 is connected is all the routers in the sub-AS 1310 (the configuration is not shown in FIG. 4) and is connected by IBGP as in FIG. All these routers are also connected to each other by IBGP, and the routers corresponding to the sub-ASs other than the external AS and the sub-AS 1310 of the own AS are EBPG-connected. The sub-AS-sub-AS and the self-AS-external AS are EBPG connections as in the case where the sub-AS 1310 in FIG. 4 has a route reflection configuration, and the peer information providing system 1 is connected via these IBGP connections and EBGP connections. By collecting BGP UPDATE of all routers in the sub-AS 1310, the routes of the local AS 13 and all external ASs are collected.

  (2) FIG. 5 shows a configuration example when collecting OSPF route information. The own AS 20 to which the peer information providing system 1 is connected is divided into a plurality of OSPF areas 21, 22, and 23. Each OSPF area is connected to a router group 211, 221, and 231 in the area, respectively. Status advertisement) monitoring nodes 210, 220, and 230 are installed. The router 5 to which the peer information providing system 1 is connected is a group of routers 211, 221, and 231 in each area via the monitoring nodes 210, 220, and 230 in each area, and receives a prefix list for each area. To do. Each monitoring node receives a Network Summary LSA (network summary LSA) in the area, and this Network Summary LSA information includes a prefix set of areas other than its own area.

  Therefore, each LSA monitoring node notifies the received Network Summary LSA to the peer information providing system 1 as it is, and the peer information providing system 1 extracts all prefix lists extracted from all Network Summary LSAs and the Network Summary of the corresponding LSA monitoring node. By taking the difference from the prefix list extracted from the LSA, it is possible to know the prefix set of the OSPF area to which the corresponding LSA monitoring node belongs. For example, when it is desired to know the prefix set of the OSPF area 21 (referred to as [prefix21]), the prefix set of the areas 23 and 21 obtained from the LSA monitoring node of the OSPF area 22 (also referred to as [prefix23, prefix21]). ) And the prefix set [prefix21, prefix22] obtained from the LSA monitoring node in the OSPF area 23 from the prefix set [prefix21, prefix22, prefix23] of all areas in the AS 20 are obtained from the OSPF area 21 to be known. [Prefix21] can be obtained by subtracting the prefix set [prefix22, prefix23] obtained from its own monitoring node and taking the difference.

  In the method (2), more detailed route information than the method (1) can be obtained for the route in the own AS, but the route of the external AS cannot be collected as in the method (1).

  In the route reception method (3), the method (1) and the method (2) are used in combination. In this case, the peer information providing system 1 connects to the BGP speaker in the same AS (own AS) by the method (1) and connects to the LSA monitoring node of the OSPF speaker by the method (2) and receives a route. Further, the router to be connected or monitored may include a router serving both as a BGP speaker and an OSPF speaker. In method (3), information that overlaps between BGP and OSPF regarding the route in its own AS will be received. However, giving priority to the route collected by OSPF based on address information etc. gives more detailed route information. Since it is obtained, it is preferable.

  Also, in any of the above methods (1) [(1-1), (1-2), (1-3)], (2), it is received directly using a routing protocol such as BGP or OSPF. How to collect route information from messages and messages such as BGP UPDATE and OSPF Network Summary LSA are temporarily stored in a file, etc., and the route information is received by the peer information providing system 1 periodically reading the file etc. Both are possible.

  Next, the route information collected from the router 5 by the route receiving unit 2 as described above is passed to the network information management unit 3, and the network information management unit 3 classifies the route information to create / update network group information. The steps will be described separately in (A) an example of network group information to be created / updated and (B) a procedure for creating / updating network group information.

  (A) First, a schematic example of network group information obtained by classifying route information by the network information management unit 3 is shown in FIG.

  As shown in FIG. 6, the network group information is segmented management information, that is, the prefix to which the address of each node (including candidate peer and requesting peer) on the network belongs is classified into segments (network range) in units of network groups. Information to be managed, and segment information listing the prefixes classified and belonging to each segment based on the segmentation management information. The ID for identifying each segment is SID (Segment ID, corresponding to PID in conventional technology P4P). In FIG. 6, five types of SIDs of SID: 1 to SID: 5 are generated as a result of receiving / updating route information. Yes. In the segment information column, prefixes belonging to each SID are listed, and [10.0.0.0/24] and [10.0.0.1/24] are assigned to SID: 1 and [10.1.0.0/ 24], [10.1.0.1/24] is [SID.3: [10.2.0.0/24], [10.2.0.1/24] is SID: 4 is [192.168.0.0/24], [192.168.1.0/ 24], [192.168.2.0/24] belong to SID: 5 with [192.168.16.0/24], [192.168.17.0/24], [192.168.18.0/24].

  In FIG. 6, the segment information column and the segmentation management information column describe examples of IPv4 addresses. However, the embodiment of the present application is not limited to the related portion of FIG. It is applicable even when using both address and IPv6 address.

  In addition, the classification of prefixes into these five types of SID: 1 to SID: 5 uses the network identification information to which the prefix included in the route information belongs using the route information from the router to each prefix. An example is shown in the segmentation management information column in FIG. This is due to the method (3) described above. The route information in the local AS is selected by giving priority to the OSPF route information of the method (2), and the route information of the external AS is selected by the method (1) according to the BGP of the EBGP connection. This is an example when UPDATEs are collected. Segmentation is performed for each OSPF area if it is a prefix of a route in its own AS, and for each AS to which the prefix belongs if it is a prefix of a route of an external AS.

  In FIG. 6, SID: 1, SID: 2, SID: 3 are examples of segmented SIDs when the local AS is divided into three OSPF areas as shown in FIG. Corresponds to the prefix in the OSPF area. Each OSPF area is identified by the IP address of the LSA monitoring node belonging to each area described in the method (2), and therefore the segmentation management information field in FIG. The IP address is described following the description of OSPF. In other words, [SID: 1 OSPF10.0.129.1] means that the prefix reached by the route to the OSPF area to which the LSA monitoring node whose IP address is [10.0.129.1] belongs and is monitored is classified as SID: 1 It means that. The same applies to [SID: 2 OSPF10.1.129.1] and [SID: 3 OSFP10.2.129.1] in the segment management information column in FIG.

  In FIG. 6, SID: 4 and SID: 5 are examples of segmented SIDs for external ASs, and are obtained by EBGP in the method (1) [(1-1), (1-2), (1-3)]. Each external AS is identified by the AS path included in the specified BGP UPDATE. If the prefix belongs to an external AS, the AS end AS (Origin AS) included in the route information from the router 5 to the prefix is the belonging AS, so the prefix is classified by the origin AS identification number. ing. That is, [SID: 4 OriginAS 100] in the segmentation management information column in the figure means that the prefix of the route whose Origin AS number of the route information is 100 in the external AS is classified as SID: 4. The same applies to [SID: 5 OriginAS 200] in the segment management information field in FIG.

  (B) Next, the route group collection method (1) [(1-1), (1-2), (1-3)], the network group information as illustrated in FIG. The procedure for creating / updating will be described with reference to FIG. In the procedure of FIG. 7, the prefix in its own AS gives a SID for each OSPF area to which it belongs (for example, SID: 4, SID5 in FIG. 6). ) A prefix belonging to an external AS is a procedure for assigning and classifying each external AS to which it belongs.

  Network group information is created / updated according to the procedure shown in FIG. When the process is started, in step S1, the peer information providing system 1 is connected to the router 5, or the route information file periodically updated by the router 5 or the like is read to read new route information. Receive. The process proceeds to step S2, where it is determined whether the received path information is an OSPF path or a BGP path. If it is OSPF, the process proceeds to step S3, and if BGP, the process proceeds to step S4.

  When the process proceeds to step S3 in the case of an OSPF route in step S2, a prefix newly obtained by route reception is assigned a SID for each OSPF area to which the prefix belongs. Refer to the network group information before route reception, and if there is already a SID corresponding to the OSPF area, a new prefix is added to the prefix list of the corresponding SID. When a new OSPF area is added and a new LSA monitoring node is provided accordingly, the corresponding SID is added, and the added SID is assigned to the prefix of the new OSPF area. Conversely, when the OSPF area is integrated / deleted, the corresponding SID and prefix are also integrated / deleted. Thus, in step 3, segmentation management information, which is a SID classification standard, and network group information as a list of prefixes belonging to each SID are created / updated. As in the case of FIG. 6, the SID of the OSPF area is associated with the IP address of the LSA monitoring node of the OSPF area.

  Note that the method for obtaining a list of prefixes belonging to each OSPF area is the same as described in Method (2). All Network Summary LSAs in the local AS (or all remaining OSPFs excluding the corresponding OSPF area in the local AS) The prefix of the corresponding OSPF area is obtained as a difference between all the OSPF prefixes in the own AS obtained from the Network Summary LSA of the area and the OSPF prefix obtained from the Network Summary LSA of the relevant OSPF area.

  When the process of step S3 is completed in this way, the process for the OSPF path in FIG. 7 ends. If there is a route reception again, the processing of FIG. 7 is started each time.

  In step S2, in the case of the BGP route, when the process proceeds to step S4, the route information is included in the BGP UPDATE (BGP route update), so that the case is classified according to the type of UPDATE. If the type of UPDATE is ANNOUNCE (route notification), the process proceeds to step S5, and if it is WITHDRAW (path cancellation), the process proceeds to step S6.

  When the UPDATE type is ANNOUNCE and the process proceeds to step S5, the prefix newly obtained in the route information in this case is included in the NLRI (Network Layer Reachability Information) in the UPDATE message. Read sequentially. The OriginAS (AS at the end of the AS path, that is, the AS to which the prefix belongs) corresponding to the read prefix is identified by a number as shown in FIG. 6 and associated with the SID, and the prefix list of the corresponding SID is updated. If there is no corresponding SID, a new SID is created in association with the Origin AS number, and the corresponding prefix is added to the list.

  Further, when the type of UPDATE is WITHDRAW and the process proceeds to step S6, in this case, the new route information is route cancellation, and the prefix to be canceled is included in the withdrawn route (cancellation route) in the UPDATE message, so these are read in order. . If there is a prefix in the list of prefixes of each SID that corresponds to the prefix read as a cancellation route, it is deleted from the list.

  As described above, in the case of the BGP route, through step S5 or step S6, the segmentation management information as the SID classification standard and the network group information as a list of prefixes belonging to the SID are created / updated, and the process ends. . If there is a route reception again, the processing of FIG. 7 is started each time.

  Note that the processing of FIG. 7 includes the time series of reception for each path reception (including the case of reading from a file if it is read from a file). The processing is started from the start according to the update recording time series of each route reception information.

  As described above, in (A) and (B), only OSPF and BGP have been described as examples of routing protocols, but IS-IS (Intermediate System to Intermediate System) can be implemented in the same manner as intra-domain routing protocols. It is.

  When the server unit 4 receives the peer list request from the user, the server unit 4 returns the peer list with priority within the same SID. Next, a process for creating a peer list for replying to the user of the server unit 4 will be described. In addition, when the user is regarded as a position to receive and return the requested peer list from the server unit 4, a Tracker (distribution control server) corresponds to a hybrid P2P such as BitTorrent (registered trademark) as the user, and Gnutella If it is pure P2P like this, it can be considered that the peer itself corresponds.

  The user sends candidate peer information returned by the distribution control server for each piece of requested content, that is, the IP addresses of a plurality of candidate peers (the number is N), the IP address of the requesting peer (user himself), and the candidate peer. A predetermined number of peers in the peer list created by selecting the server unit 4 from among all N pieces (assuming that n is a common number for each piece piece) is used as peer list request information for each request content. It transmits to the server part 4 for every piece piece.

  Upon receiving the user peer list request information, the server unit 4 searches for the affiliation SID of the requesting peer by referring to the network group information of the network information management unit 3 for the IP address of the requesting peer from the information. . When the requesting peer is included in the address space of multiple prefixes, the affiliation SID is determined by longest match (the longer prefix length is given priority). Next, the affiliation SIDs of all candidate peers are searched and determined from the IP addresses of candidate peers included in the candidate peer information in the peer list request information. Here, m is the number of candidate peers that have the same SID as the requesting peer among the candidate peers.

  Hereinafter, three methods (a), (b), and (c) for creating the peer list by selecting the predetermined number n from the candidate peers will be described.

In the first method (a), a peer list is created as follows based on the size relationship between the number m (the number of SIDs that are the same as the requesting peer among the candidate peers) and the number n (the predetermined number of peers in the peer list).
(A1) When m <n, select all m candidate peers of the same SID in the peer list. Remaining n−m are selected randomly from the remaining Nm non-identical SIDs in the peer list. (A2) When m> n Select m random peers from m candidate peers of the same SID in the peer list (a3) When m = n, select m = n candidate peers of the same SID as they are in the peer list. This is a standard method in which candidate peers with the same SID and a short network distance are preferentially selected as much as possible. In the method (a), traffic is localized, but it takes time for the content to be distributed and distributed to peers with a long network distance.

In the second method (b), as in the method (a), the case is classified according to the size relationship between the number m (the number of the same SIDs as the requesting peer among the candidate peers) and the number n (the predetermined number of peers in the peer list). To create a peer list. Then, in order to improve the point that it takes time to disperse the content in the method (a), a peer selection rate Sr (0 <= Sr <= 1, predetermined value) with a short network distance is prepared, and the following method is used. Peers with different SIDs and long network distances are selected in the peer list.
(B1) When m <n, m × Sr of the same SID candidate peers are randomly selected in the peer list and the remaining n−m × Sr are randomly selected from the N−m candidate peers of different SIDs in the peer list. (B2) When m> n, n × Sr from the peer peers of the same SID and n × (1-Sr) from the candidate peers of different SIDs are randomly selected from the n peer lists (b3) When m = n, m × Sr of the same SID candidate peers are randomly selected in the peer list. The remaining n−m × Sr are randomly selected from the N−m candidate peers of different SIDs in the peer list.

  In the third method (c), the operator predetermines priorities for a set of requesting peer SID and candidate peer SID as shown in FIG. 8 and selects a peer list using this. In this case, a peer list is created by selecting peers in descending order of priority from the pair of requesting peers and candidate peers. Method (c) replaces criterion “same SID” in method (a) [(a1), (a2), (a3)] with “priority for requesting peer among peers not yet selected in peer list” Can be implemented in the same way by replacing “the highest candidate peer SID” with the standard and replacing the number m in (a) with m ′ (where m ′ is the number of candidate peers for which priority is specified for the requesting peer). .

  In this way, in the method (c), m ′ pieces are sequentially selected from the candidate peers of the high priority SID for the requesting peer, and if m ′ is less than the number of peer lists n, the remaining candidate peers are selected. It will be chosen at random. For example, when the priority designation is shown in FIG. 8 and the requesting peer is SID1, the candidate peer of SID2 with the highest priority 20 is first selected in the peer list, and when there are no peers belonging to SID2, the next priority is given. 10 candidate SID3 peers are selected in the peer list. If all n peer lists are not selected even when there are no peers belonging to SID3, the remaining SID1 and other unspecified priority peers (peers belonging to SIDs other than SID2 and SID3) are randomly selected. The remaining peers in the peer list are elected.

DESCRIPTION OF SYMBOLS 1 ... Peer information provision system, 2 ... Path | route receiving part, 3 ... Network information management part, 4 ... Server part, 11, 12, 13, 20 ... AS, 110, 121, 122, 123, 131 ... BGP router (BGP speaker) ) 21, 22, 23 ... OSPF area, 210, 220, 230 ... LSA monitoring node, 211, 221, 231 ... OSPF router (OSPF speaker)

Claims (7)

A requesting peer that performs P2P (peer-to-peer) type communication in the network and requests the desired content transmits the ID of the desired content to the distribution control server, receives candidate peer information that holds each piece of the desired content, and A requesting peer receives a peer list request in which the candidate peer information is transmitted together with the address information of the requesting peer, and selects a peer list in which a peer to be communicated is selected from the candidate peer information based on the peer list request. A peer information providing system for P2P communication provided to
Receiving means for receiving, from a router operating in the network, route information including a route to a prefix to which the requesting peer belongs and a route to a prefix to which a candidate peer in the peer list belongs;
Classifying means for classifying a prefix included in the route information using the updated route information and associating it with a segment ID each time the route information is received and updated by the receiving means; and Peer information providing system for P2P communication.   2. The classification unit according to claim 1, wherein the classification unit classifies the prefix based on network identification information to which a prefix included in the path information belongs using the path information and associates the prefix with the segment ID. Peer information providing system for P2P communication as described. The peer information providing system for P2P type communication according to claim 1 or 2, further comprising:
By determining the segment ID of each peer as the segment ID associated with the prefix to which each peer belongs, and by selecting the peer to be communicated from the candidate peer information based on the segment ID of each peer, the peer A peer information providing system for P2P type communication, comprising: a selection means for creating a list.   The said selection means preferentially selects the candidate peer with which the segment ID of the said request | requirement peer and the segment ID of the candidate peer contained in the said candidate peer information correspond, and produces a peer list | wrist, Peer information providing system for P2P type communication described in 2.   The selection means includes priority setting means for setting a predetermined priority for a set of a segment ID of the requesting peer and a segment ID of the candidate peer, and selects the candidate peers in descending order of priority. 4. The peer information providing system for P2P communication according to claim 3, wherein the peer list is created. The router includes an OSPF router, and the receiving means receives the route information from the OSPF router via the LSA monitoring node of the OSPF area included in the network, and the route summary includes the Network Summary LSA of the OSPF area. Included,
The peer information for P2P communication according to any one of claims 1 to 5, wherein, when the prefix belongs to the OSPF area, the classification means associates the segment ID with each OSPF area. system. The router includes a BGP router, and the route information includes a BGP UPDATE issued by the BGP router,
6. The peer-to-peer for P2P communication according to claim 1, wherein, when the prefix belongs to an Origin AS in the BGP UPDATE, the classification means associates the segment ID with each Origin AS. 7. Information provision system.

Images