JP2010200026A - Traffic control method, system and program for logic network - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reduce measuring costs and overhead such as an excessive load on a network, a processing load, processing delay and the like caused by traffic measurement using a test packet between overlay nodes by eliminating the need of the traffic measurement using the test packet between the overlay nodes required for conventional overlay network communication routing. <P>SOLUTION: A logic network (overlay network 51) composed of nodes 51a-51f, a link and a management server 54 logically set on an IP network 53 is configured to estimate quality information among the overlay nodes 51a-51f using flow information measured over the IP network 53. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a technique for improving end-to-end communication quality in the Internet, and more particularly to end-to-end using a logical network such as an overlay network logically formed on an IP (Internet Protocol) network. -It relates to a technique suitable for efficiently improving end communication quality.

  The Internet, which represents an IP network, has been developed and spread to accommodate various applications. On the other hand, there is an increasing demand for QoS (Quality of Service).

  Along with this, it is an important issue to realize technology ("end-to-end QoS management technology") on the Internet to avoid end-to-end congestion and improve quality. . However, there are the following problems in realizing such a technique.

  (1) The Internet has already become a social infrastructure, and it is difficult to expand new functions at the network layer that greatly change the existing network structure.

  (2) The Internet is formed by a plurality of ASs (Autonomous Systems) having different management entities, and it is difficult to extend new functions to all ASs simultaneously.

  Under such circumstances, as an effective technique that can improve end-to-end QoS without changing a lower network layer, a QoS management technique using an overlay network described in Non-Patent Document 1, for example, has attracted attention.

  As described in Non-Patent Document 2, for example, an overlay network is a network that uses an existing link and forms a logical (virtual) link in an upper layer according to the purpose.

  The basic concept of QoS management by such an overlay network is illustrated in FIG. In FIG. 1, 1a to 1c are nodes (overlay nodes) constituting the overlay network 1 (described as “Overlay-NW” in the figure), and 2a to 2e are IP networks (described as “IP-NW” in the figure). 2, and traffic flows along a path indicated by a broken-line arrow from x to y. Further, it is assumed that there is a congested IP router on this route, and as a result, the QoS between x and y is lowered.

  At this time, using the overlay network formed by the overlay nodes 1a, 1b, and 1c, the traffic is detoured along the route (x → overlay node 1a → overlay node 1b → overlay node 1c → y) represented by the solid line arrow. If this is possible, the above congestion can be avoided.

  In fact, Non-Patent Documents 3, 4, 5, 6 and the like indicate that a large number of such detour paths exist in the real network based on actual measurements. However, these results are evaluations when an ideal communication path calculation is performed using quality information measured between all overlay nodes.

  The quality information is obtained by actively measuring between overlay nodes using test packets. Extra load on the network, processing load and processing delay associated with measurement between overlay nodes, etc. There is a problem that there is a measurement cost overhead.

  In addition, in the P2P file sharing application, a certain user downloads from a node (group) holding a desired file. In this case as well, when selecting which node to download, consider the quality between the nodes. A method of selecting a download source has been studied (for example, Non-Patent Document 7).

  It has also been reported that the download time can be shortened by considering the quality (inter-node delay) (Non-Patent Document 8).

  However, this approach also has the problem of measurement cost overhead as in the case of the overlay path control described above.

  In order to avoid the above problem, Non-Patent Document 9 discusses an approach (called P4P) in which an ISP (IP network side) and a P2P operator (overlay side) cooperate. In this technique, the ISP side discloses the topology and link load information of the IP network, and allows the P2P to select a download source node that does not concentrate the load on a specific link.

  By doing so, it is possible to select a node that can shorten the download time without the overlay side measuring itself, and the ISP side can also perform load distribution.

  However, in P4P, the flow of traffic is controlled only by controlling the selection of peers (that is, end hosts). Since end hosts behave autonomously, it is not always possible to control traffic as desired by ISPs. There was no problem.

L. Zhi and P.M. Mohapra, “QRON: QoS-aware routing in overlay net-works,” IEEE J. MoI. Select. Areas Commun. , Vol. 22, pp. 29-40, January 2004. WIDE Project, “Integrated Distributed Environment with Overlay Network,” WIDE Project Research Report, Part 17, 2002. Kamei, Kawahara, “Traffic engineering by end-host overlay network and its effectiveness,” Shingaku Sodai, BS-5-3, 2004. S. Rewaskar and J.H. Kaur, “Testing the Scalability of Overlay Infrastructures,” Proc. PAM 2004. April 2004. S. Banerjee, T .; G. Grifin and M.M. Pias, “The Interdomain Connectivity of PlanetLab Nodes,” Proc. PAM 2004, April 2004. Hiraoka, Hasegawa, Murata, “Evaluation of Overlay Routing Effectiveness Using Delay and Bandwidth Information,” IEICE Technical Report CQ2007-18, July 2007. H. Shen and Y. Zhu, “Plover: A Proactive Low-overhead File Replication Scheme for Structured P2P Systems,” IEEE ICC 2008. T.A. Karagiannis, P.A. Rodriguez, and K.R. Papagannaki, “Should Internet service providers, fair peer-assisted content distribution?” IMC 2005. H. Xie and Y.C. R. Yang, “P4P: Provider Portal for Applications,” SIGCOMM 2008.

  The problems to be solved are as follows: (1) When calculating the communication path in the overlay network, test packets are used between the overlay nodes, and the extra load on the network and between the overlay nodes There are measurement costs and overheads such as processing load and processing delay associated with measurement, and (2) When selecting a node from which a desired file is downloaded in a P2P file sharing application, In consideration of the quality between nodes (such as delay between nodes), there is a measurement cost overhead as in the case of route control in the overlay network.

  An object of the present invention is to solve these problems of the prior art and to improve end-to-end quality and traffic transfer efficiency.

  In order to achieve the above object, according to the present invention, there is provided a method for controlling traffic on a logical network configured with nodes, links, and a management server, for example, an overlay network, logically set on an IP network, (1) Quality information between overlay nodes is estimated using flow information measured in the IP network, and communication path calculation in the overlay network is performed using the flow information. Also, (2) an overlay network is formed by an overlay node (carrier overlay node) that can be controlled by a network provider such as an ISP, and the traffic status normally monitored by the IP network is fed back to the overlay network side to Link the overlay network. In addition, (3) when an end host does not exist on a route that realizes optimal traffic distribution in the IP network, the content is relayed and cached using a carrier overlay node.

  According to the present invention, in a logical network composed of nodes, links, and management servers that are logically set on an IP network, for example, an overlay network, (1) quality information between overlay nodes is transferred to the IP network. Estimated using the measured flow information, traffic measurement by test packets between overlay nodes, which is necessary for communication path calculation in the conventional overlay network, is unnecessary, and extra load on the network and between overlay nodes Measurement costs and overheads such as processing load and processing delay associated with traffic measurement using test packets can be reduced. Also, (2) an overlay network is formed by an overlay node (carrier overlay node) that can be controlled by a network provider such as an ISP, and the traffic status normally monitored by the IP network is fed back to the overlay network side to The overlay network is linked to enable traffic control in a desired form such as an ISP. In addition, (3) when there is no end host on the route that realizes optimal traffic distribution in the IP network, the content is relayed and cached using the carrier overlay node, and the end-to-end quality is improved. The efficiency of traffic transfer can be improved.

It is explanatory drawing which shows the example of a path control by the overlay network to which this invention is applied. It is a block diagram which shows the 1st network structural example to which the traffic control system of the logical network which concerns on this invention is applied. It is a flowchart which shows the example of a procedure for comprising the traffic control system of the logical network which concerns on this invention. It is a sequence diagram which shows the example of a procedure of the content delivery by the traffic control system of the logical network which concerns on this invention. It is a block diagram which shows the 2nd network structural example to which the traffic control system of the logical network which concerns on this invention is applied. It is a block diagram which shows the structural example of the management server which concerns on this invention. It is explanatory drawing which shows the operation example of the free zone | band calculation which concerns on this invention. It is a block diagram showing the structural example of the cOL node which concerns on this invention. It is a block diagram showing the example of a structure of the eOL node which concerns on this invention.

  Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings. In the overlay network 1 shown in FIG. 1, which is a logical network constructed by several overlay nodes 1a to 1c connected to the IP network 2 via IP routers 2a to 2e, the overlay nodes 1a to 1c are CPU ( The computer has a computer configuration including a central processing unit and a main memory, and performs a process related to traffic control of the logical network of the present invention by executing a programmed computer process.

  Further, as shown in FIG. 2, in this example, a management server 24 is provided, and this management server 24 also executes a programmed computer process to perform a process related to the traffic control of the logical network of the present invention. Do.

  That is, in this example, the conventional problems in the overlay network logically formed on the IP network ((1) the measurement cost and overhead are required for creating the routing table, and (2) the network provider. Network traffic cannot be controlled by ISP or the like), and (3) improvement of end-to-end quality and traffic transfer efficiency.

  In order to solve the problems (1) and (2) above, in this example, the traffic situation normally monitored by the IP network (IP network) is fed back to the overlay network side, and the IP network and the overlay network are linked. In order to achieve the above (3), a technology is provided for caching the contents together with the contents relay in the overlay node at the relay point.

  Hereinafter, the first technique for achieving the above (3) and the second and third techniques for solving the problems (1) and (2) will be described.

  In the first technique, there is an overlay network that is a logical network constructed by several overlay nodes connected to an IP network (IP network, IPNW). Among these overlay nodes, a network such as an ISP or a network carrier An overlay node prepared by the provider is called a “carrier overlay node”, and an overlay node (such as an end user PC) other than the carrier overlay node is called an “end host overlay node”.

  In an overlay network composed of these two types of overlay nodes, each carrier overlay node is used as a “representative node” to form an overlay node group, and each end host overlay node belongs to one of the groups.

  Assume that each overlay node group is logically connected to other groups. This connection relationship is called “logical topology”. The procedure so far is called “phase 1”, and the following procedure is called “phase 2”.

  Next, on the logical topology, a route for distributing the content is calculated, the content is transferred along the route, and an end host overlay node exists in the group located on the route. Then, the end host overlay node relays the content and caches the content at the same time.

  When there is no end host overlay node, the carrier overlay node performs content relaying and caching. The procedure so far is "Phase 2".

  Phase 1 will be described with reference to FIG. In FIG. 2, cOL nodes 21a, 22a, and 23a represent carrier overlay nodes, and eOL nodes 21b, 21c, and 23b to 23d represent end host overlay nodes.

  Each cOL node 21a, 22a, 23a is installed in each group 21, 22, 23. In addition, it is assumed that each of the eOL nodes 21b, 21c, and 23b to 23d belongs to any of the groups 21, 22, and 23.

  Which group belongs to, for example, it belongs to the group to which the cOL nodes 21a, 22a, and 23a closest to the eOL nodes 21b, 21c, and 23b to 23d belong. Here, “close” may be a packet delay time or may be geographically close (such as between Tokyo).

  Regarding the procedure of phase 1, the procedure from input (precondition) to output (until an overlay network for content distribution is configured) will be described in order using FIG. 3 together with FIG.

  [1-1] First, an NW provider such as an ISP or a carrier (hereinafter referred to as a carrier) arranges a server (computer) so as to connect to its own IP network. The server is arranged for each area in the IP network (for example, Tokyo, Osaka, Nagoya, etc., see FIG. 2) (step S301 in FIG. 3).

  This area corresponds to a group, and a server arranged by a carrier is a carrier overlay node cOL node.

  [1-2] Next, the carrier installs the management server 24 shown in FIG. 2 for managing the entire overlay network (overlay NW) (step S302).

  [1-3] The management server 24 manages a correspondence table between the group X23 and the IP address (IP # xxx) of the cOL node X23c belonging to the group 23 (step S303).

  [1-4] The management server 24 determines whether to logically connect the cOL nodes 21a, 22a, and 23a. That is, the logical topology is determined (step S304). Note that logical connection means that each cOL node knows the other party's IP address and is in a communicable state.

  [1-5] The management server 24 distributes logical topology information to each cOL node 21a, 22a, 23a (step S305). In the example of FIG. 2, the groups 21, 22, and 23 are connected by a full mesh.

  [1-6] Each cOL node 21a, 22a, 23a calculates an optimum route using a quality metric (for example, delay time) for each logical link in the distributed topology, and creates a route table. (Step S306). Note that how to estimate the quality metric will be described in detail after the description of the second technique.

  [1-7] On the other hand, the user who wants to receive the content distribution service (end host node of group (3) 23: eOL node 23c) inquires the management server 24 and teaches the cOL node 23a closest to the eOL node 23c. The cOL node 23a is notified of the desire to belong to the group (3) 23 and belongs (step S307).

  [1-8] The eOL node 23c obtains topology information and a route table from the cOL node 23a of the group (3) to which it belongs (step S308).

  The above is the procedure of Phase 1.

  Next, the procedure of phase 2 will be described using FIG. 4 together with FIG.

  [2-1] First, the eOL node Y21b that wants to receive content obtains the corresponding content holding list from the management server 24 (step S401 in FIG. 4).

  [2-2] The management server 24 returns a content holding list to the eOL node Y21b (step S402). The content holding list is a list of nodes (and groups to which the nodes belong) holding the content.

  [2-3] The eOL node Y21b selects a distribution source node from the content holding list. Specifically, based on the route table obtained in step S308 of phase 1, a group that can be distributed with the minimum hops as seen from itself is examined, and an eOL node or a cOL node belonging to that group is selected as a distribution source node.

  Thereafter, the eOL node Y21b transmits a content distribution request to the selected node (step S403). For example, in the example of FIGS. 2 and 4, the content distribution source node is the eOL node X 23 c of the group (3) 23.

  [2-4] The content distribution source eOL node X23c looks at its own routing table and distributes the content to the nodes belonging to the group to be transferred next.

  For example, as shown in FIG. 2, it is assumed that the optimum route for the content distribution is group (3) 23 → group (2) 22 → group (1) 21.

  If there is an eOL node participating in the group (2) 22, the content is transferred to this eOL node. If there is no eOL node in the group (2) 22, the cOL node A 22a of the group (2) 22 is designated as a relay node, the content is first transferred there (step S404), and from there, the eOL of the request source The data is transferred to the node Y21b (step S405).

  Note that the content transfer method may be a store-and-forward method in which the content itself is temporarily stored, or an on-the-fly method in which each packet or piece constituting the content is transferred to the next while being received by the cOL node A. Simple transfer is also acceptable.

  In this example, the cOL node A 22a of the group (2) 22 caches the relayed content. As a result, in the P2P using only the eOL node, traffic cannot be transferred along the optimum route via the group (2) 22, whereas in this example, the cache of the cOL node A 22a is used. This is possible.

  The process of step S403 is repeated until the content reaches the content request node, and phase 2 is realized.

  The reason why the content is cached by the cOL node (A22a) is that, for example, if an eOL node belonging to the group (2) 22 is generated in the future and the node requests the content, the original content is grouped. The reason for receiving from (3) eOL node X23c is that it is inefficient.

  On the other hand, after receiving the eOL node Y21b, the content is cached, and when there is a request for the content from the eOL node Z21c of the same group (1) 21 thereafter, the content is transferred to the eOL node. By transmitting from Y21b to the eOL node Z21c, redundant traffic transfer is avoided.

  Next, in the second technology, in the connection relationship (topology) between the overlay node groups in the first technology, as a metric of each logical link, a carrier overlay node i that is a representative node of the group i and a group j The transfer path is calculated using the communication quality (delay time, packet loss rate, available bandwidth, or a combination thereof) between the carrier overlay nodes j.

  In the third technique, the quality information between the overlay nodes in the second technique is estimated using the flow information measured in the IP network.

  Here, the flow information refers to information obtained by measurement techniques such as NetFlow and sFlow. Specifically, the number of packets and bytes are counted on a “flow” (session identified by a combination of TCP, UDP, ICMP destination, source IP address and port number) on a network device such as a router or switch. The aggregated data is transmitted to the traffic management server using a NetFlow packet, and the traffic management server analyzes the flow information collected from network devices such as routers.

  As a technique for estimating the quality using the flow information, for example, when estimating the TCP throughput corresponding to the available bandwidth, the publicly known document “R. Kawahara, T. Mori, K. Ishibashi, N. Kamiyama, and H. Yoshino, "Packet sampling TCP flow rate estimation and performance detection method," IEICE Trans. Commun., Vol. E91-B, No. 5, 19 pp. 1309-1 Mp.

  Alternatively, when estimating a packet retransmission rate in TCP corresponding to a delay time (RTT: round trip time) or a packet loss rate, for example, a public tool tcptrace (URL = http: //irg.cs) on the Internet. .. ohio.edu/software/tcptrace/).

  As described above, by using the traffic measurement information in the IP network, it is possible to reduce the traffic measurement cost by the test packet between the overlay nodes, which is necessary in the conventional overlay network.

  In addition, by feeding back the state of the IP network to the overlay network and cooperating, the traffic can be controlled in the form desired by the ISP according to the state of the IP network. For example, the information such as the free bandwidth state of the IP network is fed back as an input to the overlay, and when the route is selected on the overlay side, the route with the largest free bandwidth is selected, so that the traffic is moved to one place. Avoid concentrating.

  Next, a fourth technique according to the present invention will be described. In this fourth technology, instead of causing the carrier overlay node to perform content relaying as in the first technology, the communication path is established using only a router or switch in the IP network, and at the same time, the content cache Content caching is performed by setting a route for transferring traffic to the carrier overlay node to be performed within the router or switch.

  For example, by using a technique called “openflow (URL = http: //www.openflowswitch.org/)” disclosed on the Internet, a route of content traffic of interest is established in the IP network.

In this Openflow,
(I) In the router / switch, all the first packets from the flow not entered in the “flow table (FT)” are transferred to the controller prepared in advance.
(Ii) In the controller, specify what kind of control is to be performed on the flow, enter the result in the FT of each router,
(Iii) In each FT, the route is controlled at the flow level by writing information such as where to transfer when a packet matching the flow arrives (as in NextHop = router X). .

  Therefore, it is possible to control the route in the IP network by reflecting the route information calculated in the cOL node in the FT.

  If it is desired to cache the content in the cOL node, it is possible to receive and cache the packet at the cOL node by writing the FT so that the packet is transferred to the cOL node.

  Embodiments 1 and 2 of the configuration and processing operation of a traffic control system for a logical network according to the present invention using the first to fourth techniques will be described below.

[Example 1]
As shown in FIG. 5, the system in this embodiment has a configuration of an overlay network 51 and a lower transfer network (IP network) 53. The overlay network 51 includes carrier overlay nodes (cOL nodes) 51a to 51e, end host overlays. Nodes (eOL nodes) 51f to 51i and a management server 54 are included.

  The lower transfer network 53 includes network devices such as routers 53a to 53g, is connected to network devices 52a to 52d that terminate communication terminal devices (eOL nodes) 51f to 51i such as personal computers, and is an edge router. It is connected to the overlay network 51 via 52e-52i.

  As shown in FIG. 6 as the management server 64, the management server 54 has a traffic quality information management unit 64a, a content holding list management unit 64b, a cOL node management unit 64c, and a topology determination unit as functions for executing programmed computer processing. 64d and an area map part 64e.

  The traffic quality information management unit 64a in the management server 64 collects traffic information measured in the lower transfer network 53 and creates a traffic quality matrix between areas. Here, the area may be defined in units of prefectures such as Tokyo and Kanagawa, or may be a unit in which an edge router of a lower transfer network exists.

  The procedure for creating a traffic quality matrix between areas will be described below.

  For example, the quality d_j (throughput, delay, packet loss rate) is estimated for each flow j using the third technique described above. On the other hand, a location database that associates areas with IP addresses is prepared separately.

  With reference to the source IP address and destination IP address of flow j whose quality is estimated, the location database is used to collate the area A corresponding to the source IP address and the corresponding area B of the destination IP address.

  The flow j from the departure area A to the arrival area B is collected, and the quality d_ab from the area A to the area B is obtained by Σd_j / N_ab. Here, N_ab is the number of flows from area A to area B.

  Alternatively, the available bandwidth between areas may be calculated using each link load of the lower transfer network and route information.

  For example, as shown in FIG. 7, it is assumed that the overlay network is configured by cOL nodes A71a, B71b, and C71c, and the lower transfer network (network) is configured by routers (1) 72a to (5) 72d. In addition, it is assumed that the free bandwidths in the links in the lower transfer network are indicated by 73a to 73d.

  Here, the positions in the cOL nodes 71a to 71c and the lower transfer NW are managed by the DB 74 shown in FIG. In this example, it is assumed that areas A, B, and C correspond to cOL nodes A, B, and C, respectively.

  In addition, the DB 74 manages the presence of the router (1) 72a in the area A and the router (3) 72c in the area B in the position DB 74a. Further, in the example of FIG. 7, the route from the router (1) 72a to the router (3) 72c is configured via the router (2) 72b. In the traffic information DB 74b.

  By managing in this way, the management server 64 calculates, for example, the free bandwidth on the path from the area A to the area B as 2 Mbps.

  In addition, the content holding list management unit 64b in the management server 64 manages which content is held by which cOL node or eOL node.

  Further, the cOL node management unit 64c in the management server 64 manages a correspondence table indicating which area (= which overlay group) each cOL node belongs to. Specifically, the group ID is associated with the IP address of the cOL node installed in the group.

  Further, the topology determining unit 64d in the management server 64 determines the connection relationship (topology) of the overlay group formed by each cOL node.

  In response to the inquiry from the eOL node, the area map unit 64e in the management server 64 returns the IP address of the cOL node installed in the overlay group (area) closest to the eOL node. For this purpose, a database is managed that maps an area from the IP address of the inquiry-source eOL node.

  FIG. 8 shows a configuration example of the cOL node 81. The cOL node 81 has a topology formation management unit 81a, a route calculation unit 81b, a route table management unit 81c, a content as functions to execute programmed computer processing. A transfer unit 81d and a content cache unit 81e are provided.

  The topology formation management unit 81a grasps a connection relationship (topology) with an overlay group formed by another cOL node. The topology may be determined centrally by the management server (54, 64), or it may be determined whether or not messages are exchanged autonomously between the cOL nodes.

  The route calculation unit 81b obtains the traffic quality information d_ab between the areas A and B from the management server.

  Thereafter, in the topology formed by the topology formation management unit 81a, if the overlay group A and the overlay group B managed by the cOL node A and the cOL node B installed in the area A and the area B are connected, the logical link Let AB quality metric be d_ab.

  In this way, when a quality metric is given to each logical link on the topology, an optimum route is calculated using the metric. For example, when the delay time is used as a metric, the shortest path is calculated by the Dijkstra method.

  The content transfer unit 81d and the routing table management unit 81c perform the same operations as those in the eOL node described later.

  FIG. 9 shows a configuration example of the eOL node 91. The eOL node 91 has a cOL node selection unit 91a, a content request unit 91b, a routing table management unit 91c, a content as functions to execute programmed computer processing. A transfer unit 91d and a content cache unit 91e are provided.

  The eOL node 91 first participates in the overlay network when receiving a service, for example, when requesting reception of a certain content. Specifically, the cOL node selection unit 91a determines which cOL node belongs to. At this time, the cOL node selection unit 91a inquires of the management server, for example, and tells the cOL node closest to itself.

  The content request unit 91b makes a content reception request with reference to the content holding list. This content holding list may be centrally managed by the management server, or may be distributedly managed between cOL nodes and eOL nodes. The content request unit 91b obtains a list of cOL or eOL nodes holding content from the content holding list, specifies the nearest node from the list, and transmits a request.

  When the content transfer unit 91d receives a request from the eOL node that is the content request source, from the eOL node that relays the content, or from the cOL node, the content is managed by the routing table management unit 91c. To the adjacent node (NextHop cOL node or eOL node).

  The routing table management unit 91c acquires the routing table from the cOL node, and stores and holds it in the storage device. Also, by looking at the routing table, the cOL node Z of the overlay group Z that is NextHop is accessed, and if there is an eOL node z belonging to the cOL node, the IP address (IP # z) of the eOL node z is inquired and obtained. To do. If it does not exist, the IP address (IP # Z) of the cOL node is notified. The IP address is managed as NextHop # Z = IP # z (or IP # Z).

  The content cache unit 91e caches the content when the content is transferred.

  Next, a configuration and processing operation of the logical network traffic control system according to the present invention will be described.

[Example 2]
In [Example 1], the operations of the content transfer unit 81d and the content cache unit 81e in the cOL node 81 shown in FIG. 8 are the same as the operations of the content transfer unit 91d and the content cache unit 91e in the eOL node 91 shown in FIG. However, it may be another one shown below.

  As described in the fourth technology, a communication path is established using only a router or a switch in a lower transfer network (IP network), and at the same time, traffic is transferred to a cOL node for which content cache is desired. Content caching is performed by setting the route to be performed in the router or switch.

  Specifically, as in the first embodiment, the routing table management unit (91) in the eOL node (assumed to be the eOL node z0) obtains the routing table from the cOL node, stores it in the storage device, and holds it. Keep it.

  Further, referring to the routing table, the cOL node Z of the overlay group Z that is NextHop is accessed, and if there is an eOL node z that belongs to the cOL node, the IP address (IP # z) of the eOL node z is inquired. The IP address is managed as NextHop # Z = IP # z.

  Up to this point, the operation is the same as that of the first embodiment. In the second embodiment, when the eOL node z does not exist, the access is made to the cOL node Z2 that is one hop ahead, and if the eOL node z2 that belongs to the cOL node exists, the IP address (IP # of the eOL node z2) z2) is obtained by inquiring, and its IP address is managed as NextHop # Z = IP # z2. This operation is repeated until an eOL node is found.

  In addition, the management server is notified that there is no eOL node belonging to the cOL node Z.

  When the management server receives the notification, the optimum route for the corresponding flow, that is, the content transfer from the eOL node z0 to z2, is the route via the cOL node group Z, and the node to be cached is The fact that it is the cOL node Z is stored and managed as information in the storage device.

  On the other hand, in the router in the lower transfer network, the first packet from the flow that is not entered in the flow table (FT) is transferred to the management server.

  The management server selects what kind of control is performed on the flow. Specifically, when the optimum route and cache destination are managed for the flow as described above, the result is entered in the FT of each router, and the flow matches with the flow in each FT. When a packet to be received arrives, information that can pass through the cOL node group Z is written to control the route at the flow level.

  At the same time, it is written in the FT so that the packet is transferred to the cOL node Z, and the packet is received and cached also in the cOL node.

  As described above with reference to FIGS. 1 to 9, in this example, as a logical network composed of nodes, links, and management servers, which is logically set on the IP network 53 that is a lower transfer network. As a function for performing traffic control on the overlay network 51 and executing programmed computer processing, the overlay nodes 51a to 51i include a topology formation management unit 81a, a route calculation unit 81b, a route table management unit 81c, The content transfer unit 81d and the content cache unit 81e, or the cOL node selection unit 91a and the content request unit 91b, the routing table management unit 91c, the content transfer unit 91d, and the content cache unit 91e are provided. Quality information management unit 64a and A content holding list management unit 64b, a cOL node management unit 64c, a topology determination unit 64d, and an area map unit 64e are provided. The overlay nodes 51a to 51e obtain the traffic information of the IP network 53 from the management servers 54 and 64 by the topology formation management unit 81a, and use the obtained traffic information of the IP network by the route calculation unit 81b. A route on the overlay network 53 is selected and stored in the storage device by the route table management unit 81c.

  In addition, the overlay nodes 51a to 51i cache the relayed content by the content cache units 81e and 91e, and transmit the cache to the request source via the content transfer units 81d and 91d in response to the request for the cached content.

  The traffic quality information management unit 64a of the management servers 54 and 64 collects traffic information for each flow detected by the network devices (53a to 53g) on the IP network 53, analyzes the collected traffic for each flow, and analyzes the quality metric. And the estimation result is transmitted to the overlay nodes 51a to 51e.

  The overlay node includes a plurality of carrier nodes 51a to 51e prepared by a provider of the overlay network 51, and a plurality of end host nodes 51f to 51i other than these carrier nodes 51a to 51e. The management servers 54 and 64 have cOLs. The node management unit 64c, the topology determination unit 64d, and the area map unit 64e store and manage the carrier nodes 51a to 51e in groups for the carrier nodes 51a to 51e, and the connection state between the carrier nodes 51a to 51e. The logical topology to be shown is determined, and the determined logical topology information is transmitted to each of the carrier nodes 51a to 51e, and each of the carrier nodes 51a to 51e corresponds to each logical link in the logical topology transmitted from the management servers 54 and 64. Route calculation unit 81b More, the route selection using the traffic information.

  In addition, the management servers 54 and 64 hold a content holding list that specifies a node that holds the content for each content in the content holding list management unit 64b, and each end host node 51f to 51i is specified as an affiliation destination. The management servers 54 and 64 are inquired about the group to be processed, and the logical topology and the route information selected by the carrier nodes 51a to 51e are obtained from the carrier nodes 51a to 51e of the group specified by the management servers 54 and 64. Then, at the time of requesting content, the end host nodes 51f to 51i obtain the content holding list from the management servers 54 and 64, select the distribution source node of the requested content by referring to the content holding list, and select the content holding list. When the content distribution request is transmitted to the distribution source node, and the content distribution request is received, the end host nodes 51f to 51i refer to the route information obtained from the carrier nodes 51a to 51e of the group to which the node belongs. Select a delivery route, specify the end host nodes 51f to 51i or carrier nodes 51a to 51e belonging to the next transfer destination group in the selected delivery route as the relay node, transfer the requested content, and designate as the relay node End host nodes 51f to 51i or carry The nodes 51a to 51e relay and transfer the received content to the end host nodes 51f to 51i or carrier nodes 51a to 51e belonging to the next transfer destination group, cache the content, and request the next distribution of the content Is received, the cached content is transmitted to the request source.

  As described above, in this example, in the overlay network that is a logical network that is logically set on the IP network and includes nodes, links, and management servers, (1) quality information between overlay nodes is represented by IP. Estimated using the flow information measured in the network, the traffic measurement by the test packet between the overlay nodes, which is necessary for the calculation of the communication path in the conventional overlay network, becomes unnecessary, and the extra load on the network and overlay Measurement costs and overheads such as processing load and processing delay associated with traffic measurement by test packets between nodes can be reduced.

  Also, (2) an overlay network is formed by an overlay node (carrier overlay node) that can be controlled by a network provider such as an ISP, and the traffic status normally monitored by the IP network is fed back to the overlay network side to The overlay network is linked to enable traffic control in a desired form such as an ISP.

  In addition, (3) when there is no end host on the route that realizes optimal traffic distribution in the IP network, the content is relayed and cached using the carrier overlay node, and the end-to-end quality is improved. The efficiency of traffic transfer can be improved.

  In addition, this invention is not limited to the example demonstrated using FIGS. 1-9, In the range which does not deviate from the summary, various changes are possible. For example, in this example, the overlay network is an example of a logical network, but the present invention is not limited to this.

  In FIG. 2, one cOL node 21a, 22a, and 23a is installed in each group 21, 22, and 23. However, for example, a plurality of cOL nodes 21a, 22a, and 23a may be installed in consideration of fault tolerance.

  DESCRIPTION OF SYMBOLS 1,51: Overlay network (Overlay-NW), 1a-1c: Overlay node, 2,53: IP network (IP-NW, lower forwarding network), 2a-2e: IP router, 21,22,23: Overlay node Group, 21a, 22a, 23a, 51a to 51e: cOL node (carrier overlay node), 21b, 21c, 23b to 23d, 51f to 51i: eOL node (end host overlay node), 21d: cache, 24, 54, 64 : Management server, 52a to 52d: communication terminal device, 52e to 52i: edge router, 53a to 53g: router, 64a: traffic quality information management unit, 64b: content holding list management unit, 64c: cOL node management unit, 64d: Topology determination unit 64 : Area map part, 71a to 71c: cOL nodes A to C, 72a to 72d: Routers (1) to (5), 73a to 73d: Free bandwidth on the link, 74: DB, 74a: Location DB, 74b: Traffic information DB, 81: cOL node, 81a: topology formation management unit, 81b: route calculation unit, 81c: route table management unit, 81d: content transfer unit, 81e: content cache unit, 91: eOL node, 91a: cOL node selection unit 91b: content request unit, 91c: route table management unit, 91d: content transfer unit, 91e: content cache unit.

Claims (8)

A traffic control method on a logical network composed of nodes, links and management servers, which is logically set on an IP network,
The management server constituting the logical network collects traffic information of the IP network from the IP network,
Each node constituting the logical network obtains traffic information of the IP network from the management server, and selects a route on the logical network using the obtained traffic information of the IP network. Logical network traffic control method. A traffic control method for a logical network according to claim 1,
A traffic control method for a logical network, wherein the nodes constituting the logical network cache the relayed content and transmit the cache to a request source in response to a request for the cached content. A traffic control method for a logical network according to any one of claims 1 and 2,
The management server constituting the logical network is
Collect traffic information for each flow detected by network devices on the IP network,
A logical network traffic control method comprising: analyzing traffic collected for each flow, estimating a quality metric, and transmitting the estimation result to a node constituting the logical network. A traffic control method for a logical network according to any one of claims 1 to 3,
The nodes constituting the logical network include a plurality of carrier nodes prepared by a provider of the logical network and a plurality of end host nodes other than the carrier nodes,
The management server stores and manages each carrier node in groups for each carrier node, determines a logical topology indicating a connection state between the carrier nodes, and transmits the determined logical topology information to each carrier node. ,
A logical network traffic control method, wherein each carrier node performs route selection using the traffic information for each logical link in the logical topology transmitted from the management server. A traffic control method for a logical network according to claim 4,
The management server holds a content holding list that identifies a node that holds the content for each content,
Each end host node inquires of the management server about the group specified as the affiliation destination, obtains the logical topology and the path information selected by the carrier node from the carrier node of the group specified by the management server,
At the time of requesting content, the end host node obtains the content holding list from the management server, selects the content distribution source node to be requested by referring to the content holding list, and Send the content distribution request,
When receiving the content distribution request, the end host node refers to the route information obtained from the carrier node of the group to which it belongs, selects the distribution route, and the end host node belonging to the next transfer destination group in the selected distribution route Alternatively, transfer the requested content by designating the carrier node as a relay node,
The end host node or carrier node designated as the relay node relays the received content to the end host node or carrier node belonging to the next transfer destination group, caches the content, and stores the next content of the next time. A logical network traffic control method, comprising: transmitting a cached content to a request source when a distribution request is received. A traffic control method for a logical network according to any one of claims 1 to 5,
The logical network consists of an overlay network,
A logical network traffic control method, wherein the node comprises an overlay node. A system that performs traffic control on a logical network that is logically set up on an IP network and that includes nodes, links, and a management server,
The node and the management server comprise means for executing processing by the node and management server in the traffic control method according to any one of claims 1 to 6, as means for executing programmed computer processing. A logical network traffic control system.   The program for making a computer perform each process procedure by the node and management server in the traffic control method of the logical network in any one of Claims 1-6.

Images