JP5391777B2 - Route selection method, route selection system, and router used therefor - Google Patents

Description

  The present invention relates to a route selection method, a route selection system, and a router used therefor, and more particularly to a route selection method for selecting an optimum route when an overlay network is created on a network formed by a plurality of domains.

  A large network such as the Internet is composed of a plurality of small networks. Each of the plurality of small networks is operated by an independent policy, and is managed as a domain independently in each AS (Autonomous System) unit.

  Within each domain (intra domain), the network administrator of the organization grasps the network topology in the domain and uses a routing protocol called OSPF (Open Shortest Path First) to finely divide the network in the domain. It can be operated.

  On the other hand, in communication and interconnection between different domains (interdomain), a protocol called BGP (Broader Gateway Protocol) is used as a routing protocol between domains. This BGP provides a reachable mechanism for route selection for communication across multiple domains, but when communicating with other domains operated by these different policies, the so-called End- It becomes difficult to provide a quality of service (QoS) of to-End.

  Therefore, when creating one overlay network on the inter-domain (between multiple domains) and providing a certain service on the overlay network, the underlay network that is the foundation of this overlay network is gathered. The impact on inter-domain communication on service quality assurance must be considered. For example, Patent Documents 1 and 2 describe techniques for improving the quality of communication between interdomains.

  Currently, BGP version 4, which is generally used as an inter-domain routing protocol, provides a route search mechanism that has reachability from a transmitting terminal in a domain to a receiving terminal outside the domain. Also, routers at the boundaries of each domain can know the free bandwidth of the partner domain by exchanging information on how much traffic the domain can accept. Therefore, when it is desired to transmit traffic from inside a certain domain to outside the domain, it is possible to select a candidate next hop using the information of the empty band notified from the partner domain.

  Patent Document 1 discloses a technique for improving service quality for a service that requires real-time performance by a plurality of paths (multiple paths) from a transmission terminal to a reception terminal when communication between domains is performed. Yes. In this technology, transmission is performed with multiple paths, so if a packet loss occurs on one route, the quality of service is improved by correcting the data using the arrived packet via the remaining route. Can be done.

JP 2006-067075 A JP 2003-502941 A

  However, the above-described measurement techniques and interdomain communication methods have the following problems. The first problem is that BGP can know the available bandwidth to the other party's domain by exchanging information in inter-domain communication, but detailed route information in the other party's domain is not provided. It is not possible to confirm link bandwidth congestion that occurs in intra-domain communications. Therefore, when the empty bandwidth is less than the received notification of the empty bandwidth due to congestion in the partner domain, the service quality of the traffic deteriorates and the end-to-end service quality cannot be maintained.

  The second problem is that there is a problem in the reliability of the arrived information content because the border routers, which are routers at the domain boundaries, exchange information with different network operation policies.

  The third problem is that, as shown in Patent Document 1, when traffic is transmitted using a plurality of paths, useless traffic is always generated between the transmitting terminal and the receiving terminal, Traffic will increase. Although Patent Document 2 discloses a technique for forming a virtual network (overlay network) between a plurality of domains and connecting virtual links (tunnels) on the virtual network to select a route, A method for forming a virtual link is not specifically disclosed.

  The present invention has been made to solve the above-described problems, and an object of the present invention is to provide a route selection method and route selection system capable of improving the quality of an overlay network across inter-domains and a router used therefor Is to provide.

The route selection method according to the present invention includes:
A route selection method that forms an overlay network that is a virtual network across a plurality of domains, and performs route selection from a terminal in one domain to a terminal in another domain,
A first step of forming a plurality of overlay network planes using each formed virtual node at an edge router and a border router in each of the plurality of domains;
In each of the plurality of overlay network planes, a tunnel connection is made from an edge router (first router) of the certain domain to an edge router (second router) of the other domain to form a plurality of virtual links. The second step,
In the second router, to measure the traffic condition of the plurality of virtual links, and a third step of notifying the first router,
In the first router, a route determined by the measurement result using each traffic state that is a measurement result of the plurality of virtual links and a traffic state measured by the underlay network including the plurality of domains. And a route determined by BGP to make an optimum route selection,
It is characterized by including.

The system according to the invention comprises:
A route selection system that forms an overlay network that is a virtual network across a plurality of domains and performs route selection from a terminal in one domain to a terminal in another domain,
First means for forming a plurality of overlay network planes using virtual nodes respectively formed at an edge router and a border router in each of the plurality of domains;
In each of the plurality of overlay network planes, a tunnel connection is made from an edge router (first router) of the certain domain to an edge router (second router) of the other domain to form a plurality of virtual links. A second means,
In the second router, to measure the traffic condition of the plurality of virtual links, and third means for notifying the first router,
In the first router, a route determined by the measurement result using each traffic state that is a measurement result of the plurality of virtual links and a traffic state measured by the underlay network including the plurality of domains. And a route determined by BGP, and a fourth means for making an optimum route selection,
It is characterized by including.

The router according to the present invention
A router in a route selection system that forms an overlay network that is a virtual network across a plurality of domains and performs route selection from a terminal in one domain to a terminal in another domain,
Means for forming a plurality of overlay network planes using a virtual node formed on its own router and a virtual node respectively formed on another router;
In the case where the own router is a router of a transmitting terminal on each of the plurality of overlay network planes, a tunnel connection is made from the own router to a router (second router) on the receiving terminal side in another domain, Means for forming a virtual link of
With respect to the second router, it means for transmitting a measurement request traffic condition of the plurality of virtual links,
A route determined by the measurement result using each traffic state as a measurement result of the plurality of virtual links from the second router and a traffic state measured by the underlay network including the plurality of domains. And means for making an optimum route selection by comparing the route determined by BGP ,
It is characterized by including .

The program according to the present invention is:
A computer-readable program that controls the route selection operation of a router in a route selection system that forms an overlay network that is a virtual network across multiple domains and makes a route selection from a terminal in one domain to a terminal in another domain There,
A process of forming a plurality of overlay network planes using a virtual node formed on its own router and a virtual node formed on each of the other routers;
In the case where the own router is a router of a transmitting terminal on each of the plurality of overlay network planes, a tunnel connection is made from the own router to a router (second router) on the receiving terminal side in another domain, Forming a virtual link of
With respect to the second router, the process of transmitting the measurement request traffic condition of the plurality of virtual links,
A route determined by the measurement result using each traffic state as a measurement result of the plurality of virtual links from the second router and a traffic state measured by the underlay network including the plurality of domains. And a route determined by BGP to make an optimum route selection,
It is characterized by including .

  According to the present invention, when communicating across multiple domains, by constructing an overlay network on the inter-domain, the multiple overlay networks are created using the virtual topology and virtual links on the overlay network. Since it was created, it becomes possible to measure the end-to-end traffic state between domains with multiple virtual links (via tunnels).

  In addition to determining the route by exchanging information between domains within a domain, the traffic state is actually measured using a measurement packet at a fixed time so that the terminal on the access network side of the own domain can It is possible to select an optimum route for communication to a terminal in the domain.

It is a flowchart for demonstrating the principle of this invention. It is a figure which shows the example of the communication system between the multiple domains to which this invention is applied. It is a figure which shows the structural example of the underlay network corresponding to the system of FIG. It is a figure which shows the structural example of the overlay network obtained using the structural example of the underlay network of FIG. It is a figure which shows the example of the internal structure of the edge router and border router by embodiment of this invention. It is a figure which shows the specific example of the analysis function part 310 of FIG. It is a figure which shows the measurement aspect between an edge router and a border router. It is a figure which shows the example of the routing table in an edge router and a border router. (A) shows an example of a tunnel connection request message, and (B) shows an example of a tunnel connection response message. It is a figure which shows the transmission procedure of the tunnel connection request message of an edge router. It is a figure which shows the transmission procedure of the tunnel connection response message of an edge router. It is a figure which shows the measurement procedure of the traffic state between the edge routers between inter domains. It is a figure which shows the measurement procedure of the traffic state by the measurement packet of an edge router. It is a figure which shows the procedure at the time of reception of the traffic state of the result measured by the edge router.

  Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a flowchart for explaining the outline of the operation of the present invention. Referring to FIG. 1, an overlay network (virtual network) between a plurality of domains (interdomain) is formed (step S1).

  Here, in each network node such as a border router or an edge router, an overlay network creation function using a virtual node is installed, and an overlay network is created by this overlay network creation function.

  Explaining how to create this overlay network, each node border router and edge router is equipped with a virtual node that participates in the overlay network, and a virtual link is created between these virtual nodes using tunnel technology. As a result, a virtual topology for the overlay network is generated.

  When a virtual node in each of these network nodes joins the overlay network to create a plurality of overlay networks, a plurality of virtual paths are used by using virtual nodes and virtual links as paths between the network nodes. be able to.

  Then, on this overlay network, when selecting a route from a terminal in one domain to a terminal in another domain, a tunnel connection is made from an edge router in one domain to an edge router in another domain (step S2). The traffic state is measured (step S3). The measurement result is notified to the border router of each partner domain (step S4).

  Here, the notified measurement result is analyzed by the analysis function of the border router. That is, the parameter that is the analysis result is managed by the routing table that is the management table. In other words, the traffic state of the route determined by BGP (parameters used bandwidth, delay, packet loss rate), which is a measurement result by the underlay network for the destination, and control via a tunnel (virtual link) on multiple overlay networks The traffic state of the route is managed with the management table.

  The traffic state of the route determined by BGP, which is the measurement result by the underlay network, is measured in parallel with the measurement of the traffic state via the tunnel on the overlay network.

  Then, based on the traffic state management table of each route, it is determined whether the route selected by the BGP of the underlay network is the optimum route (step S5). If it is not the optimum route, the optimum route is selected from the traffic state management table for the prefix.

  Here, the prefix will be briefly described. In order for the router in the core of the network to record all the routing information, an enormous memory capacity is required, and the route search time becomes long. However, aggregating the routing tables reduces the amount of routing information. One method for this is to look at the first bit (Prefix) of an IP (Internet Protocol) address (when the address is viewed from left to right, the left is the upper bit), and consider multiple IP addresses as shared (same) bits. Address aggregation.

  Further, the destination next hop and the output port can be determined by matching the IP address to be transmitted with the prefix address recorded in the routing table (Longest Prefix Match: LPM). The first bit of the address used in this way is called an address prefix.

  FIG. 2 is a system schematic diagram to which the present invention is applied, and shows an example of a large-scale network such as the Internet that performs communication across a plurality of domains. In FIG. 2, the network is composed of a plurality of domains AS 1 to 4, and AS 1 includes ER (edge router) 12 and 14 and BR (border router) 13 and 15. Assume that AS2 has ER21, 24 and BR22, 23, AS3 has ER31, 34 and BR32, 33, and AS4 has ER43, 44 and BR41, 42.

  Here, the transmitting terminal existing in the prefix 11.0.0.0/8 (11) of the subnetwork in AS1 is the receiving terminal in the prefix 42.0.0.0/8 (45) of the subnetwork existing in AS4 outside AS1. Suppose you want to send traffic to. In the initial setting of the entire network, the traffic from the AS 1 can be calculated from the BGP protocol by notifying and recording each other's communication route and the bandwidth with the BGP protocol for inter-domain communication. Is transmitted from domain to domain, such as ER12 → BR13 → BR32 → BR33 → BR41 → ER44 (indicated by a thick line).

  Here, referring to FIG. 3 and FIG. 4, an example is shown in which an overlay network is formed with respect to a large-scale network composed of a plurality of domains AS1 to AS4 shown in FIG. 3 and 4, the same parts as those in FIG. 2 are denoted by the same reference numerals. First, referring to FIG. 3, FIG. 3 shows an example of an underlay network, and it is assumed that virtual nodes 200 are mounted on routers such as ER and BR of each domain. As an example of the virtual node 200, it is assumed that there are at least three indicated by a circle, a rectangle, and a triangle, and 1 to 3 are mounted for each router.

  FIG. 4 shows an example in which three virtual networks, overlay networks 201 to 203, are formed by selecting one virtual node of each of these nodes (circles, rectangles, and triangles). Yes. That is, each border router BR and edge router ER in each domain is equipped with at least one virtual node 200, and each of these virtual nodes belongs to a virtual network. On the virtual network, the border routers are connected to each other through a tunnel, and a virtual network topology is formed (201 to 203).

  On these virtual networks, there is no concept between domains, and each border router and edge router belong to the same virtual network, are managed in a certain address space, and a virtual topology is constructed. By measuring traffic between edge routers using such a configuration and sharing the measurement results on other overlay networks, traffic from the edge routers in its own domain to edge routers in other domains is measured. It becomes possible to do.

  FIG. 5 is a diagram illustrating an example of an internal basic configuration of the edge router (indicated by 301), but the same applies to the border router. First, the underlay forwarding function unit 305 and the underlay routing function unit 309 process an underlay routing protocol such as OSPF or BGP, and transfer an underlay packet.

  Each of the virtual node unit (A) 302 and the virtual node unit (B) 303 has an overlay routing function unit 308 and an overlay forwarding function unit 307 of the virtual network in order to participate in the virtual network. The overlay routing function unit 308 processes the routing protocol of the virtual network, and the overlay forwarding function unit 307 processes the forwarding protocol of the virtual network.

  The encapsulation function unit 306 has an encapsulation function for tunnel technology. This is because a virtual link is created between virtual networks using tunnel technology. The measurement function unit 304 monitors the tunnel state, obtains information such as a use band, a delay, and a packet loss, which are parameters of the tunnel state, and supplies the information to the analysis function unit 310. Details of the analysis function unit 310 will be described with reference to FIG.

  FIG. 6 shows an example of the configuration of the analysis function unit 310. This analysis function unit includes a tunnel connection request unit 406 that makes a tunnel connection request when the own router wants to establish a tunnel connection, and a tunnel connection response unit 409 that responds when a tunnel connection is requested from the other party. Have. These two functional units are used when determining the partner to be measured between the edge routers.

  In addition, this analysis function unit has a mechanism for determining a measurement target partner between edge routers by a measurement packet reception processing unit 408 and a tunnel connection request message transmission processing unit 407.

  Further, the analysis function unit includes a measurement result receiving unit 402 that receives a measurement result from the measurement packet reception processing unit 408, a measurement result transmission unit 403 that transmits the measurement result to the other party, and route selection of the overlay network. And an overlay network route selection unit 404 and a tunnel state calculation unit 405 for calculating a tunnel state.

  FIG. 7 is a diagram showing a mechanism of a measurement request and measurement result notification between the transmission / reception edge routers 12 and 44 shown in FIG. A measurement packet between the edge routers is generated and transmitted to the other edge router (503). Further, when addressed to itself via a tunnel from the partner, the tunnel state calculation unit 405 calculates the traffic state from the partner's edge to itself and transmits the result to the partner's edge router by the measurement result transmission unit 403 (504) ).

  Here, referring to FIG. 10, FIG. 10 is a flowchart showing a procedure when a tunnel connection is requested between edge routers in the inter-domain. First, when sending traffic from a terminal in its own domain to a terminal in another domain, the edge router in its own domain sends traffic to the terminal address in the other domain of that partner, and then takes charge of the prefix address of that partner. Search for edge routers in the domain.

  In order to search for a partner edge router, a tunnel connection request message 711 shown in FIG. 9A and a tunnel connection response message 712 shown in FIG. Exchange information.

  First, the edge router in its own domain checks on the routing table whether or not it already has the address of the edge router of the traffic to be transmitted (steps 801 and 802). If not, a tunnel connection request message is created (step 803), and this message is transmitted to the edge router in the other domain in charge of the partner's prefix address (step 804).

  As shown in FIG. 9A, the tunnel connection request message 711 includes an IP header 701, an identification header 702 for identifying the message, an IP address 703 of the transmitted edge router, a prefix address 704 in charge, an edge The AS number 705 to which the router belongs is included.

  FIG. 11 shows a flowchart of a procedure for processing the tunnel connection request message 711 by an edge router in another domain. When receiving the tunnel connection request message, the edge router determines from the identification header 702 that it is a tunnel connection request message (steps 811 and 812), and sets the IP address of the edge router of the partner domain included in the message. Extraction is performed (steps 813 and 814), and a tunnel connection response message 712 is returned to the IP address of the edge router that transmitted the message (step 815).

Incidentally, in FIG. 9 (B) is a diagram showing an example of a tunnel connection response message 712, an IP header 706, an identification header 707, a receiving edge router IP address 708, a charge prefix address 709, receiving the edge router AS address 710.

  Further, the edge router that has received the tunnel connection response message makes a connection with the edge router in the other domain of the partner through the tunnel, and continues the traffic state between the transmission edge router and the reception edge router across the domains.

  If it is not a tunnel connection request in step 812, the received packet is transferred to the access network (step 816). In step 813, if it is not addressed to the prefix that it is in charge of, the next hop transfer of the received message is performed (step 817).

  FIG. 12 is a flowchart of a procedure for measuring a traffic state between inter-domain edge routers. FIG. 8 shows an example of a routing table in the edge router. The routing table is divided into two, one is a routing information area 601 for an existing inter-domain (such as BGP), and the other is a routing information area on a virtual network generated by an overlay network topology ( 602).

  The feature of the present invention is that the next hop address 605 calculated by the routing protocol between existing BGP inter-domains and the next hop on the virtual network topology to which the virtual node in the edge router belongs. The address 606 is separated.

  When using BGP, which is an existing inter-domain routing protocol, from the local domain to a prefix in another domain, the next hop is a measurement packet for one route determined by the BGP protocol (605). By transmitting, the traffic state is measured.

  When using a virtual network generated by an overlay network topology, the virtual node belongs to the virtual network, and by mounting multiple virtual nodes on the network node, even one network node has multiple network topologies. Even with the same destination prefix (604), there are multiple next hops or routes that pass through (606). It is possible to measure the traffic state of a plurality of routes to a destination prefix at.

  First, measurement between edge routers is started using the address of the partner edge router (step 821), and the header of general traffic is encapsulated by the destination address of the partner (step 822). Here, when using BGP, which is an existing inter-domain routing protocol, the next hop calculated by BGP is selected (step 823), and the measurement packet is transferred to the next hop determined by BGP (step 823). 824).

  When a virtual network generated by the overlay network topology is used, following step 822, an overlay network is selected (step 825), and a measurement packet encapsulated on the selected overlay network is transmitted. (Step 826). At this time, if there are a plurality of surfaces (a plurality of overlay networks), steps 825 and 826 are repeated.

  FIG. 13 is a flowchart showing the measurement of the traffic state by the measurement packet of the edge router. When the edge router receives the measurement packet (step 831), the information on the empty state, the used bandwidth, the delay, and the packet loss of the traffic state of the route through which the measurement packet has passed is received (step 832). calculate. The calculated result is returned to the edge router that has transmitted the measurement packet (step 833), so that the traffic state information of the route can be shared with each other.

  FIG. 14 is a flowchart at the time of reception of the traffic state as a result of measurement by the edge router. When the edge router that has transmitted the measurement packet receives the measurement result (step 841), the traffic state of the route that is the measurement result is recorded in the routing table in units of the next hop (step 842), and the other party's edge Evaluate the best route to the router.

  Therefore, the traffic state information measured in a plurality of routes is recorded in units of the next hop of the routing table in the edge router, and information on the empty bandwidth, used bandwidth, delay, and packet loss of the traffic state of each route (step 832), the edge router in its own domain can select the optimum route for the destination prefix in the other domain.

  As described above, the router has a function capable of creating one or more virtual nodes, and each virtual node belongs to an overlay network across interdomains. In addition, each virtual node is provided with a unique routing protocol and forwarding protocol to support services that provide each overlay network. Each border router has a global IP address and is equipped with a function for tunnel connection with each other.

  When traffic is transmitted from the own domain to the outside of the domain, first, the border router that transmits traffic transmits the traffic using the next hop determined by BGP.

  Here, a border router for traffic transmission participating in the virtual network uses a virtual link connected by a tunnel of the virtual network to transmit traffic to a traffic reception prefix. An edge router that receives traffic in another domain measures the state of the traffic and notifies the traffic transmission edge router and border router of the traffic state in that section.

  Therefore, by using multiple routes on the overlay network for a prefix address of one destination, the traffic state to that destination can be measured, and the optimal route can be selected by combining with the routing table in the router. It can be done.

  To better understand the above-described invention, it will be further described with reference to FIGS. In the large-scale network such as the Internet shown in FIG. 2, as described above, the transmission terminal existing in the prefix 11.0.0.0/8 (11) of the subnetwork in the domain AS1 exists in the AS4 outside the AS1. Consider a case in which it is desired to transmit traffic to a receiving terminal in the subnetwork prefix 42.0.0.0/8 (45).

  In the existing network, as described above, by using BGP, the communication path and the bandwidth of each other are notified between the domains, and the information is recorded in the routing table as the path information. The Specifically, the domain AS4 announces (notifies) the AS number of its own AS4 and the prefix of the subnetwork in charge to the adjacent domain AS3. Further, the domain AS3 adds its own AS number to the BGP message notified from the AS4, and notifies the adjacent AS1.

  Therefore, the AS 1 can grasp the route information from the AS 1 to the AS 4 based on the BGP message propagated from the AS 3. In particular, the voter router BR13 in the AS1 can determine that the next hop of the route to the terminal 45 is the border router in the AS3 by exchanging message information of the BGP protocol. The traffic is transmitted from the AS1 to the receiving terminal 45 of the AS4 through the route indicated by the thick line in FIG.

  However, in the inter-domain communication by BGP in FIG. 2, even if the route indicated by the bold line as described above is calculated, for example, when a congestion state occurs in the network in the AS 3, the route of the border routers BR32 to 33 is The communication quality is lowered due to the congestion, and as a result, the communication quality from the transmission terminal 11 to the reception terminal 45 is lowered. This congestion state in AS3 cannot be known by AS1 unless notified from AS3.

  Therefore, in the present invention, the network configuration shown in FIGS. 3 and 4 is used. A virtual node 200 is mounted on each border router BG and edge router ER in each domain, and the virtual node belongs to a virtual network. Further, on the virtual network, the border routers are connected to each other through a tunnel (virtual link) to form a virtual topology of the virtual network (201 to 203).

  On this virtual network, a virtual topology is constructed by virtual links connected between the virtual nodes, so that in addition to the next hop route calculated by the above-described BGP, a virtual topology visible on the virtual network is displayed. It can be used. By using such advantages, traffic measurement is performed for each virtual link of the overlay networks 201 to 203 as each virtual network in addition to the route determined by BGP between the edge routers.

  Therefore, by utilizing the characteristics of multiple overlay networks, multiple routes can be generated for a certain prefix, and traffic status can be measured for each virtual route. The optimum route is selected by comparing the route to be determined and the route determined by BGP.

  It is obvious that the control operation for route selection in each of the routers can be configured such that the operation procedure is stored in advance in a recording medium as a program and is read and executed by a computer.

1-4 Domain 12, 14, 21, 24, 31, 34, 43, 44 Edge router (ER)
13, 15, 22, 23, 32, 33, 41, 42 Border router (BR)
200, 302, 303 Virtual node (virtual router)
201-203 Overlay network (virtual network)
304 Measurement function unit 305 Underlay forwarding unit 306 Encapsulation function unit 307 Overlay forwarding unit 308 Overlay routing unit 309 Underlay routing unit 310 Analysis function unit 402 Measurement result reception unit 403 Measurement result transmission unit 404 Overlay network route selection unit 405 Tunnel state Calculation unit 406 Tunnel connection request unit 407 Tunnel connection request message transmission processing unit 408 Measurement packet reception unit 409 Tunnel connection response unit

Claims (18)

A route selection method that forms an overlay network that is a virtual network across a plurality of domains, and performs route selection from a terminal in one domain to a terminal in another domain,
A first step of forming a plurality of overlay network planes using each formed virtual node at an edge router and a border router in each of the plurality of domains;
In each of the plurality of overlay network planes, a tunnel connection is made from an edge router (first router) of the certain domain to an edge router (second router) of the other domain to form a plurality of virtual links. The second step,
In the second router, to measure the traffic condition of the plurality of virtual links, and a third step of notifying the first router,
In the first router, a route determined by the measurement result using each traffic state that is a measurement result of the plurality of virtual links and a traffic state measured by the underlay network including the plurality of domains. And a route determined by BGP to make an optimum route selection,
A method comprising the steps of:
  The method according to claim 1, wherein each of the virtual nodes belongs to a virtual network on the plurality of domains to construct a virtual network topology.   The method according to claim 2, wherein a virtual link that is the virtual network topology is generated by tunnel connection between the virtual nodes.   In the second step, the first router sends a tunnel connection request message to the second router, and in response to receiving the message, the second router sends the first router to the first router. 4. The method according to claim 1, wherein a tunnel connection response message is transmitted to make the tunnel connection.   In the third step, in response to the reception of the tunnel connection response message, the first router requests the second router to measure the traffic state via the tunnel, and the second router 5. The method according to claim 4, wherein the measurement result is notified to the first router.   In the fourth step, a traffic state as a measurement result from the second router is compared with a traffic state determined in the underlay network including the plurality of domains, and the route is determined according to the comparison result. 6. The method of claim 5, wherein the selection is made.   In the third step, when transmitting a measurement packet for measuring the traffic state from the first router to a destination address prefix, an existing underlay network consisting of the plurality of domains The method according to claim 1, wherein the measurement packet is transmitted to a next hop address calculated by a routing protocol to measure a traffic state to the second router. .   In the third step, when transmitting a measurement packet for measuring the traffic state from the first router to a destination address prefix, using a virtual topology and a virtual link on the virtual network The method according to claim 1, wherein the measurement packet is transmitted to the next hop address to measure the traffic state to the second router. A route selection system that forms an overlay network that is a virtual network across a plurality of domains and performs route selection from a terminal in one domain to a terminal in another domain,
First means for forming a plurality of overlay network planes using virtual nodes respectively formed at an edge router and a border router in each of the plurality of domains;
In each of the plurality of overlay network planes, a tunnel connection is made from an edge router (first router) of the certain domain to an edge router (second router) of the other domain to form a plurality of virtual links. A second means,
In the second router, to measure the traffic condition of the plurality of virtual links, and third means for notifying the first router,
In the first router, a route determined by the measurement result using each traffic state that is a measurement result of the plurality of virtual links and a traffic state measured by the underlay network including the plurality of domains. And a route determined by BGP, and a fourth means for making an optimum route selection,
A system characterized by including.
  The system according to claim 9, wherein each of the virtual nodes belongs to a virtual network on the plurality of domains and constructs a virtual network topology.   The system according to claim 10, wherein a virtual link that is the virtual network topology is generated by tunnel connection between the virtual nodes.   In the second means, the first router sends a tunnel connection request message to the second router, and in response to reception of the message, the second router sends a message to the first router. The system according to claim 9, wherein the system is configured to transmit a tunnel connection response message to establish the tunnel connection.   In the third means, the first router requests the second router to measure the traffic state via the tunnel in response to receiving the tunnel connection response message, and the second router The system according to claim 12, wherein the system is configured to notify the measurement result to the second router.   The fourth means compares a traffic state as a measurement result from the second router with a traffic state determined in the underlay network composed of the plurality of domains, and determines the route according to the comparison result. The system of claim 13, wherein the system is configured to make a selection.   When the third means transmits a measurement packet for measuring the traffic state from the first router to a destination address prefix, an existing underlay network composed of the plurality of domains is used. 15. The system according to claim 9, wherein the measurement packet is transmitted to a next hop address calculated by a routing protocol to measure a traffic state to the second router. .   The third means uses a virtual topology and a virtual link on the virtual network when transmitting a measurement packet for measuring the traffic state from the first router to a destination address prefix. The system according to claim 9, wherein the measurement packet is transmitted to the next hop address to measure the traffic state to the second router. A router in a route selection system that forms an overlay network that is a virtual network across a plurality of domains and performs route selection from a terminal in one domain to a terminal in another domain,
Means for forming a plurality of overlay network planes using a virtual node formed on its own router and a virtual node respectively formed on another router;
In the case where the own router is a router of a transmitting terminal on each of the plurality of overlay network planes, a tunnel connection is made from the own router to a router (second router) on the receiving terminal side in another domain, Means for forming a virtual link of
With respect to the second router, it means for transmitting a measurement request traffic condition of the plurality of virtual links,
A route determined by the measurement result using each traffic state as a measurement result of the plurality of virtual links from the second router and a traffic state measured by the underlay network including the plurality of domains. And means for making an optimum route selection by comparing the route determined by BGP ,
A router characterized by including:
A computer-readable program that controls the route selection operation of a router in a route selection system that forms an overlay network that is a virtual network across multiple domains and makes a route selection from a terminal in one domain to a terminal in another domain There,
A process of forming a plurality of overlay network planes using a virtual node formed on its own router and a virtual node formed on each of the other routers;
In the case where the own router is a router of a transmitting terminal on each of the plurality of overlay network planes, a tunnel connection is made from the own router to a router (second router) on the receiving terminal side in another domain, Forming a virtual link of
With respect to the second router, the process of transmitting the measurement request traffic condition of the plurality of virtual links,
A route determined by the measurement result using each traffic state as a measurement result of the plurality of virtual links from the second router and a traffic state measured by the underlay network including the plurality of domains. And a route determined by BGP to make an optimum route selection,
The program characterized by including.

Images