KR100405806B1 - LST determining method for multicasting in MPLS network - Google Patents

Abstract

The present invention relates to a traffic path determination method for multicasting in a multi-protocol label switching (MPLS) network, wherein two tree paths are determined for a single point-to-multipoint connection and the two paths are determined. It is possible to optimally divide traffic demands, but to apply source-based tree path determination method instead of destination-based tree path. Also, traffic requiring multicast arrives at ingress as entrance. When the MPLS tree path is determined, an on-line method is applied, and not only the traffic flow rate (yi) but also the traffic demand amount (d) required for transmission in the entire network. A traffic route determination method for multicasting in a multi-protocol label switching network that provides a balanced distribution of traffic load. .

Description

Traffic route determination method for multicasting in multi-protocol label switching network {LST determining method for multicasting in MPLS network}

The present invention relates to a traffic path determination method for multicasting in a multi-protocol label switching network. More specifically, the present invention relates to the VOD and AOD with the growth of the current Internet Protocol (IP) based network. , Multi-casting for a variety of multimedia information such as video conferencing and Internet movies is required, since the multi-casting multimedia information is usually required to have a high quality of service (QoS), in the present invention has a high degree of service quality We propose a method for determining two tree paths for multicast traffic transmission, which is a method of effectively multicasting multicast traffic with high quality of service, which is mainly an on-line multicast tree path. This applies to the determination of the label switch tree path (LST).

Multi-Protocol Label Switching (MPLS) is designed to ensure the quality of service (QoS) for traffic transmitted over Internet Protocol (IP) networks and to improve the efficiency of network resources. It is a technology that has been in the spotlight recently as a traffic engineering method.

MPLS traffic engineering determines label switched paths (LSPs) in a variety of ways to better accommodate the varying QoS requirements of users' IP packets. This LSP is a point-to-point path connecting a single egress from a single ingress.

Meanwhile, with the rapid growth of IP-based networks, demands for various multimedia services such as VoIP, VOD, AOD, Internet movies, and video conferencing are increasing. In many of these multimedia services, a point-to-multipoint type of service is required. Providing a point-to-multipoint type of multimedia service using an LSP connecting point-to-point is a method for network resources. Poor in terms of utilization and / or quality of service management. The reason for this is that providing a multicasting service using the LSP is to transmit multicast traffic by unicast.

Currently, the Internet Engineering Task Force (IETF) has internet drafts for MPLS label distribution using the existing multicast IP protocols PIM-SM and PIM-DM. Standardization work is in progress for multicast.

However, since the tree path setting used in the existing multicasting protocol is a destination-based method, it is difficult to properly reflect the intention of the service provider or the user. In addition, the conventional method for transmitting multicast traffic by a single tree path does not use network resources efficiently, resulting in a degradation of quality.

In addition, since the tree path determination method which considers the current network condition only takes account of the traffic volume for the traffic that is currently required to be transmitted, the tree path determination method does not balance the traffic load over the network. there is a problem.

Accordingly, the present invention has been made to solve the existing problems in the technical background as described above, and an object thereof is to provide a tree path determination method for multicast traffic in an IP network to which MPLS is applied. For example, it is possible to determine two tree paths for a single point-to-multipoint connection and to optimally allocate traffic demand to the two tree paths, but not the destination-based tree path but the source-based tree. In addition, it is possible to apply the path determination method and to apply the on-line method of determining the MPLS tree path when the multicast-required traffic arrives at the ingress as the entrance. Balance traffic loads across the network, taking into account not only the status but also the traffic demand that needs to be transmitted It is an object of the present invention to provide a traffic route determination method for multicasting in a multi-protocol label switching network.

1 is a diagram illustrating an example of a data communication network to which the MPLS scheme is applied;

2 is a diagram illustrating an example of an MPLS network having two tree paths proposed by the present invention.

3 is a block diagram of an apparatus to which the method of the present invention is applied;

4 is a flowchart illustrating a traffic path determination method for multicasting in a multi-protocol label switching network according to an embodiment of the present invention.

※ Explanation of code for main part of drawing

31 input / output interface 32: switching structure,

33: memory 34: control unit

100: Data communication network with multi-protocol label switching

110: Ingress Label Edge Router

120: egress label edge router

130,131,132,133: Label Switch Router

In order to achieve the above object, a traffic path determination method for multicasting in a multi-protocol label switching network according to the present invention is a multi-protocol label switching (MPLS) scheme in a data communication network to which multicast traffic is applied. In the method of determining the point-to-multipoint tree path (LST) for transmission, the amount of traffic demanded to be transmitted by the current capacity of the current traffic can be transmitted among the nodes of the network and the entire link between the nodes. A first step of creating a first network from which links less than are removed; Calculating a distance information value for each link of the first network based on the traffic flow amount currently used on the link and the information on the traffic demand amount; A third step of determining a first point-to-multipoint traffic path based on the calculated distance information value; A fourth step of determining whether to determine whether to de-duplicate the traffic path; And if it is determined that the non-redundant principle is not observed, a second network is created from the first network, wherein the traffic demand is the flow rate for each link of the first point-to-multipoint traffic path, A first step of creating a third network consisting of only links having a spare capacity of '0' or more among the links of the created second network; and if it is determined to comply with the non-redundant principle, the first point-to-point in the first network. A fifth step consisting of a second step of creating a fourth network from which a link passing through the multi-point traffic path is removed, and a fifth network consisting of only links having a spare capacity greater than or equal to '0' among the links of the generated fourth network; ; A sixth step of determining a second point-to-multipoint traffic path based on the generated distance information value of each link of the third or fifth network among the calculated distance information values; And a seventh step of separately allocating and transmitting the traffic demands to be transmitted on the determined first point-to-multipoint traffic path and the second point-to-multipoint traffic path.

In the second step, a delay function of each corresponding link is obtained based on the information on the traffic flow amount and the traffic demand amount, and the derivative value of the delay function is used as the distance information value.

Hereinafter, a traffic path determination method for multicasting in a multi-protocol label switching network according to a preferred embodiment of the present invention will be described in detail with reference to the accompanying drawings.

In the present invention, a path for multicast is a tree connecting point-to-multipoint, which will be replaced by the term treepath, and likewise referred to below. Or alternative tree path) means a path determined to transmit traffic in place of the tree path when a failure occurs in the tree path (or the main tree path). In other words, in the MPLS network, the tree path refers to the connection path from a single ingress router to a plurality of egress routers, and defines and uses the new term label switched tree (LST). Shall be.

1 is a diagram illustrating an example of a data communication network to which the MPLS scheme is applied. The network 100 includes a plurality of routers, and serves as an inlet of the network 100 to any data traffic among the plurality of routers. The router is called an Ingress Label Edge Router (LER) 110, and the router serving as an exit is called an Egress Label Edge Router 120. One or more routers serving as nodes of an intermediate path between the ingress LER 110 and the egress LER 120 are referred to as a label switching router (LSR) 130 and the ingress LER 110. The traffic path determined from the LSR 130 to the egress LER 120 is referred to as a label switch path (LSP1, LSP2). The LSP is an ingress LER 110. Determined for incoming traffic , In the figure, for example LSP1 means a path determined by 110 → 132 → 133 → 120, LSP2 refers to a path determined by 110 → 131 → 120.

FIG. 2 is a diagram illustrating an example of an MPLS network having two tree paths proposed by the present invention. As shown in the figure, Egress1 and Egress2 are generated from one ingress. If multicast traffic requiring multicast to (Egress2) and Egress3 arrives at the ingress as the source, two tree paths determined for transmission of the arrived multicast traffic at that ingress, i.e. And a first label switched tree path LST1 and a second label switched tree path LST2.

That is, the path of 'ingress → LSR2 → LSR5 → egress 1,2,3' represented by a thick solid line represents the first tree path LST1, and the ingress → LSR3 → LSR5,6 → represented by a thick dotted line. The paths LSR4 → Egress 1, 2, 3 represent the second tree path LST2. These two tree paths LST1 and LST2 are source-based tree paths that are determined by the ingress when there is a need for multicast traffic.

For example, suppose that 10 multicast traffic transmissions are required from ingress to egress1, egress2 and egress3. The LST1 and LST2 may be determined in the ingress, and the two tree paths LST1 and LST2 may be allocated to the multicast traffic having a total of 10 in a 6: 4 ratio. In FIG. 2, LST1 and LST2 are determined without guaranteeing link non-redundant property, but in the present invention, two tree paths with non-redundant property are determined.

FIG. 3 is a schematic block diagram of an apparatus for determining LSP and / or LST in the LER 110 and 120 of FIG. 1 and / or the ingress of FIG. 2, and is an input / output function serving as a data communication interface with an external device. An interface 31, a physical switching structure 32, a memory 33 for storing a routing table, a routing table stored in the memory 33 are created and the switching structure 32 is controlled according to the created and stored routing table. Control unit 34 for controlling the input / output of the traffic via the input / output interface 31, and the routing table stored in the memory 33 under the control of the control unit 34. Such LSPs and / or LSTs are also determined at the time of writing. Therefore, the present invention described below may be applied to the control unit 34 and implemented, and in particular, the present invention may be applied to the ingress LER as shown in FIG.

FIG. 4 is a flowchart illustrating a traffic path determination method for multicasting in a multi-protocol label switching network according to an embodiment of the present invention. It is assumed that the following information is already known to an ingress LER of an MPLS network. . Physical topology, cost function (e.g., delay) for the network, initial and spare capacity of the network link, and traffic demand for traffic that currently requires multicasting (i.e. traffic demand to be transmitted). Routers on an MPLS network (LSRs) also assume that they know the information known to the ingress LER and have a copy function for label and traffic.

In addition, the symbols and definitions used to describe the present invention are as follows.

Let G be the network at which you want to determine LSP or LST for new traffic, and G = (m, n), m is the number of links, n is the number of nodes, and yi is the link i for network traffic. Allocated flows (i = 1, ..., m), (S, MG) are point-to-multipoints for multicast traffic demand, where S is the source MG is a multicast group. d is the demand for multicast traffic between S and MG, fi (xi) is the cost incurred when link i flow is xi (i = 1, ..., m), and fi '(xi) is xi flow Is the incremental cost per unit flow, where the differential function of fi (xi) (i = 1, ..., m), dist (i) is the distance of link i (i = 1, ..., m), G (z) is a network in which the link free capacity of the link is less than z in network G, tk denotes the amount of traffic allocated to the k-th tree path for LST, where k is 1 or 2.

In FIG. 4, first, all of the nodes constituting the network G to which the MPLS scheme is applied and all the links between the nodes, remove all the links whose surplus capacity of the currently transportable link (hereinafter referred to as Ci) is less than the traffic demand d to be transmitted at present. In step S401, a new network G (d) is generated.

Subsequently, the distance information value dist (i) is obtained based on the traffic flow amount yi currently used and the information on the traffic demand d for each link i of the network satisfying G (d), wherein yi and d For example, when information is used as a variable of the delay function fi (xi) as a cost function for each link, the distance information value dist (i) is obtained by differentiating the delay function fi (yi + d) based on Equation 1 below. Since it is defined as a result value of the function fi '(yi + / d), the corresponding distance information value is obtained using Equation 1 and yi and d value information (S402).

In Equation 1, i denotes a corresponding link (i = 1, ..., m) (m = total number of links), dist (i) is a distance information value of the link i, tci is the initial of the link i The capacity, yi is the amount of traffic flow in use on link i, and ci is the current spare capacity of link i, and a relationship of "ci = tci-yi" is established. In addition, if the delay function representing the delay on each link i is fi (xi), a formula of “fi (xi) = xi / (tci-xi)” is established, where xi is the amount of traffic passing through link i. If the derivative function of the delay function fi (xi) is fi '(xi), the expression " fi' (xi) = ci / (tci-xi) 2 " is established. Therefore, Equation 1 shows a result of arranging the delay function fi (xi) by putting yi + d instead of the variable xi in the derivative function fi '(xi).

That is, fi (xi) is a function of occurrence cost when the flow rate of link i is xi. In this embodiment, a delay function is used as an example of the cost function as described above, and fi '(xi) is the flow rate of link i. Is the increase cost function per unit flow, where is the derivative of the delay function.

For reference, the cost function is an example of using a delay function in the present invention, but is not limited thereto. The cost function may be a function for obtaining distance information on a link based on information on yi and d.

Next, based on the distance information value calculated for each link i of the network G (d) as above, the first point-to-multipoint tree path in that network G (d), namely (S, MG The tree path is calculated to be used as the path of LST1. If LST1 is found in plural numbers, only one LST is determined based on the principle of determining the tree path for the link having relatively small calculated distance information. In the present invention, it is particularly preferable that the sum of the distances by the tree paths is appropriately minimized within a predetermined range. The reason for this is that the more you find the optimal tree path, the greater the computational amount. Therefore, the setting range for the sum of the distances is set to a range found by the proper tree path in the range in consideration of the calculation amount (S403).

Subsequently, it is determined whether or not to determine LST2 as the second tree path to be determined in the next step to be disjoint with LST1 as the first tree path (S404).

When it is determined in step S404 that the LST1 and the LST2 may overlap each other, a network G1 having d as the flow amount for each link of the first tree path LST1 obtained in the step S403 is generated from the network G (d). A network G1 (0) having only a link having a spare capacity of '0' or more among the generated links of the network G1 is generated (S405-1).

On the other hand, if it is determined in step S404 that LST2 is non-duplicated to LST1, a new network G2 is generated in the network G (d) from which all the links passed by the LST1 are removed, and a spare capacity among the links of the generated network G2 A network G2 (0) having only a link equal to or greater than '0' is generated (S405-2).

Next, the distance information value dist (i) is obtained based on the traffic flow amount yi currently used and the information on the traffic demand amount d for each link i of the generated network G1 (0) or network G2 (0). Since the method is the same as in step S402 and the networks G1 (0) and G2 (0) are included in the network G (d), the distance information values already obtained in step S402 are used as they are. In addition, a tree path for the second point-to-multipoint tree path, that is, (S, MG) is obtained from the network G1 (0) or G2 (0) and used as the tree path LST2. Here, the method for obtaining the tree path LST2 is the same as the step S403 (S406).

Lastly, the traffic demand amount d to be transmitted is distributed to t1 and t2 on the determined LST1 and LST2, respectively, to be transmitted, and more specifically, based on Equation 2, Equation 3, and Equation 4 below. Obtaining xi while varying the ratio of t1 and t2 to a predetermined ratio, substituting the obtained xi value into Equation 5 according to the ratio of t1 to t2 when the resultant value of Equation 5 becomes the minimum, The total traffic amount d to be transmitted on the LST1 and the LST2 is distributed and allocated (S407).

, i = 1, ..., P _km (k = 1or2)]

,

Here, Equation 2 indicates that the multicast traffic demand d is satisfied by two tree paths LST1 and LST2, t1 and t2 are traffic amounts to be allocated to LST1 and LST2, and d is the total amount of traffic to be transmitted. , δ _i ¹ is ^{1 if} the tree path LST1 crosses the link i and 0 otherwise, δ _i ² is 1 if the tree path LST2 crosses the link i and 0, t _km is the tree path LST1 (k = 1 Or the number of links of the tree path LST2 (k = 2), ci represents the current spare capacity of link i. In addition, the solution of Equation 6, that is, t1 is the amount of traffic to be allocated to LST1 among d and t2 is the amount of traffic to be allocated to LST2 among d.

As described in detail above, the traffic path determining method for multicasting in a multi-protocol label switching network according to the present invention determines two tree paths for a single point-to-multipoint connection and the two tree paths. It is possible to allocate and allocate traffic demands optimally, but to apply the source-based tree path determination method and to determine the MPLS tree path when the multicast-required traffic arrives ingress. It is applied to on-line method, and it distributes traffic load in the whole network in consideration of not only current network condition but also traffic demand that needs to be transmitted.

Therefore, according to the present invention as described above, the network resource can be efficiently utilized from the service provider's perspective. In other words, by efficiently distributing the traffic to be transmitted throughout the network while reducing the transmission cost incurred for the traffic transmission, the network resource is efficiently used. In addition, from the user's point of view, the service can be provided as an improved service quality. In addition, by making non-redundant two MPLS tree paths determined for multicasting, it is possible to determine a robust MPLS tree path for network failure. That is, multicast traffic transmission by one MPLS tree path causes a lot of traffic to be lost when a problem occurs in the tree path, but according to the method of the present invention using two MPLS tree paths, only some traffic is lost and the remaining traffic is healthy. There is an advantage.

Claims (4)

In a data communication network to which multi-protocol label switching (MPLS) is applied, a source-based determination method of a point-to-multipoint tree path (LST) for transmission of multicast traffic, A first step of creating a first network from which links of a node and nodes of the communication network are removed, wherein links whose current traffic transmission capacity is less than a traffic demand to be transmitted at present are removed; Calculating a distance information value for each link of the first network based on the traffic flow amount currently used on the link and the information on the traffic demand amount; A third step of determining a first point-to-multipoint traffic path based on the calculated distance information value; A fourth step of determining whether to determine whether to de-duplicate the traffic path; And if it is determined that the non-redundant principle is not observed, a second network is created from the first network, wherein the traffic demand is the flow rate for each link of the first point-to-multipoint traffic path, A first step of creating a third network consisting of only links having a spare capacity of '0' or more among the links of the created second network; and if it is determined to comply with the non-redundant principle, the first point-to-point in the first network. A fifth step consisting of a second step of creating a fourth network from which a link passing through the multi-point traffic path is removed, and a fifth network consisting of only links having a spare capacity greater than or equal to '0' among the links of the generated fourth network; ; A sixth step of determining a second point-to-multipoint traffic path based on the generated distance information value of each link of the third or fifth network among the calculated distance information values; And And a seventh step of separately allocating and transmitting the traffic demands to be transmitted on the determined first point-to-multipoint traffic path and the second point-to-multipoint traffic path. A method for determining traffic paths for multicasting in a protocol label switching network. The method of claim 1, In the second step, the delay function of each corresponding link is obtained based on the traffic flow amount and the information on the traffic demand amount, and the derivative value of the delay function is used as the distance information value. Traffic route determination method for multicasting in network. The method of claim 2, The distance information value is " ", Where i denotes the corresponding link (i = 1, ...., m) (m = the total number of links), and dist (i) is the distance information value of the link i, tci Is the initial capacity of link i, yi is the traffic flow in use on link i, ci is the current spare capacity of link i, and a relationship of "ci = tci-yi" is established, and d is the traffic demand to be transmitted now. A traffic route determination method for multicasting in a multi-protocol label switching network. The method of claim 1, In the seventh step, xi is calculated while varying the ratio of t1 and t2 to a predetermined ratio based on Equation (1), Equation (2) and Equation (3) below. ) And on the first point-to-multipoint traffic path and the second point-to-multipoint traffic path, depending on the ratio of t1 to t2 when the result of equation (4) becomes the minimum. Wherein the traffic demands to be transmitted are distributed and distributed, respectively, wherein t1 and t2 are traffic amounts to be allocated to the first and second point-to-multipoint traffic paths, respectively, and d is the transmission. The total amount of traffic to be taken, δ _i ¹ is ^{1 if} the first point-to-multipoint traffic path crosses link i; otherwise 0, δ _i ² is the second point-to-multipoint traffic path to link i. 1 past and 0 otherwise, P _km is the first point-to-multipoint traffic path (k = 1) or the second point-to-many The number of links in the mid-point traffic path (k = 2), ci denotes the current redundant capacity of link i, wherein the traffic path determination method for multicasting in a multi-protocol label switching network. --- (One) , i = 1, ..., P _km (k = 1or2) --- (2) , --- (3) --- (4)