KR20030026399A - Routing method for distributing load in IP network with link state routing protocol - Google Patents

Abstract

PURPOSE: A method of performing routing for load distribution of an IP network using a link state routing protocol is provided to transmit packets according to flow capacitance ratios of each path by using a maximal flow algorithm in a sub network, thereby distributing load in a network having different links. CONSTITUTION: A start point router(14) and routers(15,16) included in a sub network(10) are connected to a 155M rank link(21) and a 622M rank link(22), respectively. An end point router(17) and the routers(15,16) are connected to a 100M rank link(24) and a 1G rank link(25), respectively. The routers(15,16) are connected to a 45M rank link(23) to each other. If all routers(11-17) use a link state routing protocol, connected states of each link and bandwidth information are the same. When subscribers received in the routers(11,12,13) of the sub network(10) transceive data by having access to an Internet network(20), the subscribers have to pass through the sub network(10) composed of the routers(14-17). To perform load distribution as satisfying bandwidth amounts in the sub network(10), the start point router(14) and the end point router(17) are designated, and a maximal flow of the start point router(14) and the end point router(17) is calculated in the sub network(10).

Description

Routing method for distributing load in IP network with link state routing protocol

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a routing method in an IP (internet protocol) network. In an IP network composed of links having different bandwidths, the maximum flow in a sub-network including a path from an origin router to an endpoint router ( The present invention relates to a routing method that performs load balancing in a network composed of different links by delivering packets according to the flow capacity ratio of each path using a maximal flow algorithm.

In order to deliver a packet to an destination in an IP network, a packet must be routed (routed) and forwarded to each router hop by hop. Each router performs routing and packet forwarding based on the destination IP address of the IP packet header.

Routing in an IP network is divided into static routing in which an administrator sets routing information in each router and dynamic routing in which a routing protocol is set using a routing protocol.

The larger the network, the more difficult it is for administrators to use static routing for all network information on each router.

Dynamic routing is divided into internal routing protocol for exchanging routing information inside autonomous system and external routing protocol for exchanging routing information with autonomous system of other companies. Here, an autonomous system is a group of networks using the same routing policy by being managed and controlled as belonging to an organization in terms of management on the Internet.

The internal routing protocol maintains network information to dynamically perform packet routing, and selects an optimal path by performing routing by destination using a different metric for each protocol.

The distance vector protocol establishes a route by each router maintaining the best hop number for each destination. In the routing information protocol (RIP), the hop number to the destination is measured based on a measure. The Internet gateway routing protocol (IGRP) is based on a complex measure of bandwidth, delay reliability, loading, and maximum transmit unit for the link to the destination.

The link state protocol shares the link information of the network between routers and establishes a path by performing a shortest path first for each destination in each router. In an open shortest path first (OSPF), the bandwidth of the link to the destination is set. Use a measure based on

Each router creates a routing table by selecting a path having an optimal measure for each destination by a routing protocol enabled in a network. In this case, when there are a plurality of identical cost paths (equal cost paths) for each destination, all the corresponding paths are included in the routing table.

The router performs load balancing and redundancy using this same scale path.

In general, a packet processing procedure performed in a router is as follows.

First, when a packet is input to an input interface, the router copies and transmits the packet to a buffer in a routing processor where routing takes place to select the path of the packet.

The routing processing device looks up the routing table using the destination IP address in the IP packet header.

If the network information corresponding to the destination IP exists in the routing table, the packet is copied to the output interface for transmission to the destination.

Next, send a packet from the output interface to the next hop router. In this case, when there are a plurality of same-scale paths in the routing table, the router performs a load balancing for each path by selecting each path round-robin for a packet inputted with the same destination network information as a destination. do.

This results in packet delivery at the same rate for multiple identical measure paths.

Therefore, in a network using an internal routing protocol based on bandwidth, a link having the same bandwidth must be installed to install a plurality of links in a specific section for load balancing and redundancy.

However, if a plurality of links are installed in the same band, it is easy to economically set up excessive links. For example, if you need 650M band between two points, using 622M link and 45M link can meet the requirements, but you need to install two 622M links because you need to install the same bandwidth. have.

Here, a method of satisfying the requirements by installing 15 45M-class links is rarely used because it is less feasible due to the complexity of management.

If load balancing is performed by using the 622M link and the 45M link and designating the same scale path for the two links, the router delivers the packets at the same rate, so congestion occurs at the 45M link, resulting in 650M. Only traffic within 90M of the requirement is handled normally, and more than half packet loss may occur for the remaining packets.

That is, since the path selection method for efficiently performing load balancing for links having different bandwidths is not implemented, there is a problem in that an excessive link more than the bandwidth required between any two points must be installed.

An object of the present invention for solving the above problems is to designate a starting point router and an endpoint router including paths for which load balancing is desired, and to perform load balancing by efficiently using a link using a maximum flow algorithm in a designated subnetwork. To provide a routing method.

1 is a structural diagram of an IP network composed of links having different bandwidths as a preferred embodiment of the present invention.

2A to 2D are structural diagrams illustrating a maximum flow calculation procedure between a start point router and an end point router used in the present invention.

<Description of Symbols for Major Parts of Drawings>

10: subnetwork

11-13: Router Accepted in Subnetwork

14: starting point router

15, 16: Router included in sub network

17: endpoint router

20: Internet network

21-25: Link

In the IP network using the link state routing protocol of the present invention for achieving the above object, a routing method for links having different bandwidths may include a starting point router and an endpoint router among routers of a subnetwork including paths for load balancing. Setting up; Calculating a maximum flow capacity for all paths from the starting router to the ending router; Calculating a ratio of flow capacity between all paths from all routers in the sub-networks to endpoint routers; And forwarding a packet according to a ratio of flow capacity in each router.

The above and other objects and features and advantages of the present invention will become more apparent from the following detailed description taken in conjunction with the accompanying drawings.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a structural diagram of an IP network composed of links having different bandwidths as a preferred embodiment of the present invention.

The IP network shown in FIG. 1 includes a sub-network 10 that desires load balancing composed of a plurality of routers 14-17, and routers 11, 12, 13 accommodated in the sub-network 10, The sub network 10 includes a starting point router 14 of the sub network 10 to which routers 11, 12, and 13 accommodated in the sub network 10 and a sub network 10 to which the internet network 20 is connected. It is configured to include an endpoint router 17.

Here, the routers 15 and 16 included in the starting point router 14 and the sub network 10 are connected to the 155M class link 21 and the 622M class link 22, respectively, and the endpoint router 17 and the sub network are connected. Routers 15 and 16 included in 10 are connected to the 100M class link 24 and the 1G class link 25, respectively.

Routers 15 and 16 included in the sub network 10 are connected by a 45M class link 23.

Assuming that all routers 11-17 shown in FIG. 1 use the link state routing protocol, the link state and bandwidth information of each link are kept the same.

Subscribers accommodated in the routers 11, 12, and 13 accommodated in the sub-network 10 must pass through the sub-network 10 composed of routers 14-17 in order to access and transmit data to the Internet 20. .

In order to perform load balancing while efficiently satisfying the bandwidth requirements in the sub-network 10, a starting point router 14 and an endpoint router 17 are designated, and in the sub-network 10, a starting point router 14 and an endpoint router 17 Calculate the maximum flow between

2A to 2D are structural diagrams illustrating a maximum flow calculation procedure between the start point router 14 and the end point router 17 used in the present invention.

Find any path from the starting point router 14 to the ending point router 17 that has a strictly positive flow capacity. Here, if no path is found, the maximum flow is already obtained.

First, in FIG. 2A, a possible flow capacity (the smallest capacity of the links constituting the path) 622M in a path formed of two links 22 and 25 is determined and a decrease in the flow capacity of the path by the corresponding flow capacity is made. Let's do it. That is, as shown in FIG. 2B, the link 22 has no available flow capacity (0M), and the other link 25 has the remaining flow capacity (378M) minus the flow capacity (622M) of the link 22. It becomes the available flow capacity.

Next, in FIG. 2B, the possible flow capacity 100M in the path formed by the two links 21 and 24 is determined, and the flow capacity of this path is reduced by the corresponding flow capacity. That is, as shown in FIG. 2C, the link 24 has no available flow capacity (0M), and the other link 21 has the remaining flow capacity (55M) minus the flow capacity (100M) of the link 24. This is the available flow capacity.

In FIG. 2C, since there are no available flow capacities of the two links 22, 24, the possible flow capacities 45M are determined in the path formed by the three links 21, 23, 25, and the corresponding flow capacities are determined by the corresponding flow capacities. Decreases in flow capacity. That is, as shown in FIG. 2D, the link 23 has no available flow capacity (0M), the link 21 has 10M, and the link 25 has an available flow capacity of 333M.

This operation is repeated until there is no path from the starting router 14 to the ending router 17 with a strictly positive flow capacity.

Thus, packets can be delivered through the sub-network 10 by the total flow capacity 767M that can be delivered in all paths from the starting point router 14 to the endpoint router 17.

Here, in the case of using a general routing method in which load balancing is performed only in the same scale path, load balancing is not performed in the sub network 10, and only up to 622M can be transmitted through the sub network 10.

A routing table at each router 14-16 using a routing method that provides maximum flow load balancing according to the present invention is shown in the following Tables 1, 2 and 3, respectively.

Table 1 Routing table from the starting point router 14 to the endpoint router 17

Next link Measure ratio(%) 22 622 81 21 145 19

In the routing table of the starting point router 14 of Table 1, when the next link is the link 21 connected to the router 15, a path through two links 21 and 24 and three links 21 and 23 are used. , 25, etc., so that there are a plurality of paths, such as a path through the link 21, that is, the sum of the available capacities of all paths using the corresponding link 21, that is, the usable capacity (100M) and three in the path through the two links (21, 24) The sum 145M of the available capacity 45M in the path through the links 21, 23, 25 is used as a measure.

Table 2 Routing Table from Router 15 to Endpoint Router 17

Next link Measure ratio(%) 24 100 69 23 45 31

Table 3 Routing Table from Router 16 to Endpoint Router 17

Next link Measure ratio(%) 25 667 100

Each router 14-16 uses the available capacity of each path as a scale so as to set a path capable of maximum flow from the starting point router 14 of the subnetwork 10 to the ending point router 17.

Since each router included in the path performs the routing and forwarding process of packets by hop by hop, only the next link to be forwarded is shown in the routing table without including the entire path.

When the packets to be forwarded through the starting point router 14 and passing through the ending point router 17 arrive at the starting point router 14, each router checks the source IP address information and the destination IP address information for each packet. Then, it is checked that the packet is to be delivered to the end point router 17 through the start point router 14.

At this time, when there are a plurality of next links to the endpoint router 17, packet forwarding to the next link is distributed to the plurality of links by distributing packets according to the ratio of the link-specific measures of the routing table, that is, the ratio of available capacity. )do.

On the other hand, packets that do not pass through the end point router 17 through the start point router 14 and have different origins or different destinations perform packet processing procedures in the general router.

The routing method of the present invention described above can be equally applied to various Internet network structures.

As described above, in the routing method according to the present invention, in an IP network using a link state routing protocol and configured with different bandwidths, a sub network including paths desired for load balancing is designated, and Since the routing information is configured to allow maximum flow and load balancing is possible for unequal cost paths having different measures, it is necessary to provide different bandwidths suitable for the bandwidth required for link distribution considering load balancing and redundancy. Even if a link is installed, the load balancing routing can be efficiently performed.

In addition, a preferred embodiment of the present invention is for the purpose of illustration, those skilled in the art will be able to various modifications, changes, substitutions and additions through the spirit and scope of the appended claims, such modifications and changes are the following claims It should be seen as belonging to a range.

Claims (6)

A routing method of an IP network that controls load balancing of a subnetwork including links having different bandwidths using a link state routing protocol, Setting a starting point router and an endpoint router among routers of the sub-network including the paths desired for load balancing, calculating a maximum flow capacity for all paths between the starting point router and the endpoint router, And calculating a ratio of the flow capacity between all the paths from all the routers to the endpoint routers of the sub-networks, and forwarding packets according to the ratio of the flow capacity of each path. The method of claim 1, The calculation of the maximum flow capacity for all paths from the starting router to the endpoint router, Finding an arbitrary path from the starting router to the ending router with a strictly positive flow capacity; Determining possible flow capacity in any path established in the step of finding any path; And And reducing the determined flow capacity by the flow capacity of the set any path. The method of claim 2, Each step above, Routing method characterized in that iteratively repeats until the path from the starting router to the endpoint router having a strictly positive flow capacity disappears. The method of claim 2, The flow capacity possible in any of the established paths is Routing method characterized in that the smallest flow capacity of the capacity of the link constituting the path. The method of claim 2, The maximum flow capacity for all paths from the starting router to the endpoint router is And the sum of all flow capacities determined in the step of determining possible flow capacities for all the paths set in the step of finding an arbitrary path from the starting point router to the endpoint router. The method of claim 1, The method of calculating the ratio of the flow capacity between all the routers from all routers of the sub-network to the endpoint router, Calculating flow capacity of all paths from each router to the endpoint router; And And calculating a ratio of flow capacity between all paths by using the calculated flow capacity of all paths.