KR20010004871A - Quality of Service(QoS) based generic shortest path search method - Google Patents

Abstract

PURPOSE: A method for searching a path supporting the quality of service(QoS) is provided to make each node decide whether routing information has to be retransmitted, not to retransmit the routing information to a node through which the routing information passes, and to transmit the routing information to an adjacent node to make each intermediate node transmit a path request only, which is appropriate for the QoS, to the adjacent node, and to delete or store routing information which arrives after transmitting the routing request without re-transmitting the arriving routing information, so as to prevent inefficiency of a CPU and a bandwidth. CONSTITUTION: An algorithm initializes a set for presenting connection information of nodes within every path layer on a multi-level network and initializes the quality of service(QoS) value, to search a shortest path of many routines to many destination nodes from a single source node. The algorithm designates a starting point for a path searching to call a path searcher, and waits for a result message from the path searcher, to distribute the path searching. When the path searching is performed according to the calling, the path searching is not re-passed in a layer area through which the path searching passes. When the path searching is completed, a result message is transmitted to a called target.

Description

Quality of Service (QoS) based generic shortest path search method}

The present invention relates to a path search method for routing based on a quality of service (hereinafter referred to as QoS) required for transmitting multimedia information on various information communication networks.

In general, a route search method may be described. First, a distance vector algorithm is used. The distance vector algorithm is a path search algorithm that is widely used in the Internet.

A typical example using the above scheme is the Routing Information Protocol (RIP).

The simple operation of the distance vector algorithm is described below. First, the distance value of each node connected to each network is 0 in the path search table, and the distance of other nodes in the network is zero. The value is initialized to infinity.

Then, in each step to find the route, find the shortest distance from the current node to another node and register it in the router table N. The formula for finding the shortest distance is:

D (v) ← Min [D (v), D (w) + L (w, v)]

Where D (v) is the distance from the current node to the destination node, and L (w, v) is the distance from node w to node v.

The shortest distance information found above is transmitted to neighboring nodes connected to it, and this procedure is continued until information about all nodes connected to each node is collected from each node.

Then, when no path change information is generated at each node and no information is being transmitted at each connected link, the process recognizes that all path information has been collected and terminates the path information collection process. .

The biggest advantage of the distance vector algorithm described above is that it is simple. As shown in the above execution procedure, the algorithm is very simple and easy to implement.

On the other hand, the disadvantage of the distance vector algorithm is that it consumes a relatively large amount of time and network bandwidth to collect all the path information at each node. In addition, the disadvantage of the Ping-Pong problem is that in the environment where the network state changes frequently, the information transmitted from each node is transferred to the neighboring node. In other words, there is no function to select an optimal QoS at each intermediate node.

Second, there is a link-state algorithm. The link-state protocol is a path search algorithm, also called "Shortest Path First," which includes a distributed database model and provides the shortest path. We use the "Dijkstra's Shortest Path" algorithm to find it.

In describing the simple operation of the link-state algorithm, each local node provides its current link state to other nodes (LSA: Link-State Advertisement).

The status information of each node includes the operating interface of each node, the transmission time when the information is transmitted through the corresponding interface, and information on where the interface is connected.

The state information of each node is transmitted to other nodes using a floating function, and the transmitted information is stored in each node's link-state database.

Each node uses the stored information to find the shortest distance from each node to other nodes using the "Dijkstra's Shortest Path" algorithm and stores it in the path search table.

Link-state algorithms are known as "Open Shortest Path First (OSPF)" for Transmission Control Protocol / Internet Protocol (TCP / IP) and Private Network-to-Network Interface for Asynchronous Transfer Mode (ATM) networks. PNNI) and the like.

The disadvantage of the link-state algorithm described above is that it is more complicated to implement than the distance-vector algorithm, and wastes a lot of bandwidth by periodically providing link state information (LSA).

In addition, the link-state algorithm collects all the path search information to maintain an optimal path suitable for the initial stage. Therefore, when various multimedia services are required in different nodes, various QoS must be satisfied for each service. However, there is a point that it is not suitable as a link-state or distance vector in which paths between all nodes are already established.

Third, the Dynamic Source Routing (DSR) algorithm is an algorithm proposed by the "MANET" group of the Internet Engineering Task Force (IETF). It is not a general path searching algorithm in wired and wireless environments, but a special wireless mobile station without a base station. Path search algorithm for "ad-hoc" network. Also, no QoS support for multimedia services is considered here.

Referring to the operation of the dynamic source path search algorithm, the source node, which wants to transmit data, sends a path search setting request to a destination node to all adjacent nodes.

Each intermediate node to which the route request is sent transmits only the route request that arrives the earliest for the same route request, and then deletes the route information that has arrived without retransmitting.

The route request arriving at the destination records the list of nodes that have passed through to the source node, and is then passed to the source node.

When the path is established, all data packets transmitted contain intermediate node list information that the packet must pass, and the packets are transmitted in this order.

Such a dynamic source algorithm has a disadvantage of operating only in a special network environment called an "ad-hoc" wireless network, and there is a disadvantage in that there is no consideration for a multimedia service because it transmits a routing packet with a short time delay unconditionally.

Disclosure of Invention The present invention aims at implementing an optimal path search algorithm supporting QoS for a multimedia service in a general wired / wireless communication network in view of the above-described conventional problems. Including the route information passed between the destination nodes, each node decides whether or not to retransmit the route information so that the route information is not retransmitted, and transmits the route information to neighboring nodes. Only the route request best suited to the QoS required for the request is transmitted to the neighbor node, and the route information arrived thereafter is deleted or stored as necessary without retransmission, thereby saving CPU and bandwidth due to retransmission. The purpose is to prevent waste.

In addition, each route information arriving at the destination may be one or more. In this case, if the destination node selects a desired route and sends a reply message back to the original node, the reply message is registered in the message. By arriving at the raw node via an intermediate node, the route is established. When all nodes on the network send a route request message to find all routes at the same time, each node does not send a reply message. It aims to shorten the path setting time by setting the path by recognizing the path setting request message transmitted from the node as a reply.

In addition, various path information arriving at the destination node can be used as a back-up path or to increase the transmission speed using multi-path depending on the purpose. An object of the present invention is to provide a path searching method.

1 is a flowchart illustrating a multipath discovery (MCDGSPA) process supporting QoS according to the present invention.

FIG. 2 is a flowchart illustrating a path searching process performed when a starting point corresponds to a source node in FIG. 1. FIG.

FIG. 3 is a flowchart illustrating a path searching process performed when a starting point does not correspond to a source node in FIG. 1.

4 is a flowchart illustrating a disjointed multi-path discovery process (DMDGASP) that supports QoS in a first embodiment according to the present invention.

5 is a flowchart illustrating a detailed process of the path searching process of FIG. 4.

FIG. 6 is a flowchart illustrating a multipath discovery (MCHDGSPA) process for supporting QoS in a two-level network according to a second embodiment of the present invention.

7 is a flowchart illustrating a disjointed multi-path discovery process (DMHDGSPA) supporting QoS in a two-level network according to a third embodiment of the present invention.

The route searching method in the information communication network according to the present invention to operate as described above is located in all the routes existing on the network to search the shortest route of the multiple routes occurring from the single source node to the multiple destination nodes. A first step of initializing a set and a quality of service (QoS) value indicating connection information of nodes;

A second process of calling a route search means by designating a starting point for the route search and receiving a result message from the route search means to distribute the route search when the initialization is completed;

And performing a path search according to the call, and then transmitting a result message to the call target.

In addition, the path search method in the information communication network according to the present invention to operate as described above, to search all the paths satisfying the reference QoS value among the plurality of paths that can occur from a single source node to a single destination node. A first step of initializing an aggregation and a quality of service (QoS) value representing connection information of nodes in all paths existing on the network;

A second step of calling the route search means by specifying a starting point to search for the route and a reference QoS value to distribute the route search when the initialization is completed, and receiving a result message from the route search means;

And performing a path search according to the call, deleting the remaining paths except one of the paths in which duplicate nodes exist among the found plurality of paths, and then transmitting a result message to the call target. It is done.

In addition, the path search method in the information communication network according to the present invention to operate as described above, the shortest path of the plurality of paths occurring from a single source node to a plurality of destination nodes on a network of the network layer is multilevel. A first step of initializing an aggregate and a quality of service (QoS) value indicating connection information of nodes in a layer in all paths existing on the multi-level network to

A second process of calling a route search means by designating a starting point for the route search and receiving a result message from the route search means to distribute the route search when the initialization is completed;

When performing the route search according to the call, the region of the layer that has once passed through is not rerouted, and when the route search is performed, a third process of transmitting a result message to the call target.

In addition, the path search method in the information communication network according to the present invention to operate as described above, the reference QoS value of a plurality of paths that can occur from a single source node to a single destination node on a network where the network layer is multi-level A first step of initializing a set and a quality of service (QoS) value representing connection information of nodes in all paths existing on the multi-level network to search all paths satisfying the s;

A second step of calling the route search means by specifying a starting point to search for the route and a reference QoS value to distribute the route search when the initialization is completed, and receiving a result message from the route search means;

When performing a path search according to the call, the area of a layer that has been one-time does not go through again, and deletes the remaining paths except one of the paths where duplicate nodes exist among a plurality of searched paths, and then outputs the result message. It characterized in that it comprises a third process of transmitting to.

The above objects, features and advantages will become more apparent from the following detailed description taken in conjunction with the accompanying drawings. Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

First, in order to implement a search algorithm for providing a multimedia service in a wired or wireless network, the following matters should be considered.

First, the most important problem in designing path search algorithms for multimedia services is that in most cases there may be at least one individual optimal path for one parameter, but there is no optimal path for all parameters. to be. For example, when two parameters (bandwidth and delay) are considered in path search, in most cases the optimal path for bandwidth and the optimal path for delay will be different. As a result, a path search supporting Qos requires a decisive protocol capable of determining whether a set path is optimal for a desired QoS.

Second, mobile terminals operating in a wireless network can move from time to time. As the mobile terminal changes the network topology from time to time, and the static routing method for establishing a communication path before the path search request is inappropriate, an adaptive routing protocol should be considered.

Third, since a terminal operating in a wireless network has less storage space and less computing power than a router operating in a wired network, a path search protocol requiring complicated calculations is inappropriate, and a simple path search protocol is required. do.

Fourth, a terminal operating in a wireless network must operate a distributed path discovery protocol operable in a distributed environment. In a wireless network environment, a protocol for establishing a path search path using a central control node frequently changes data due to the fact that a wireless terminal can move from time to time. In addition, since the terminal path information for each time period for managing the changed data is required and consumes a lot of network bandwidth, it is not suitable in a wireless network environment. As a result, for wireless mobile communication, a distributed path discovery protocol is required in which a router maintains minimal information for path discovery by itself.

Hereinafter, the shortest path search algorithm implemented in the present invention will be described.

1 is a flowchart illustrating a multipath search algorithm (MCDGSPA) supporting QoS according to the present invention.

This algorithm is used when multiple destinations are specified from one source for sending a message. If there is a duplicate node in the middle of a path that arrives at multiple destinations, the message is transmitted to the same path.

Initialize the total QoS value for any node s with a value of 'MAXQ', meaning that this node is a node that is not passed, a value with information about that node, and a value for the next node to be connected to this node. Initializes to an empty set (S1).

For reference, the algorithm and variable definitions are as follows.

typedef struct {Qos tqos; Oset set; } Route;

struct {Qos tqos; Oset set;} node [s];

Oset nextset [s];

Qos linkqos [s] [N];

Route route;

node [s] .tqos = MAXQ;

node [s] .set = {@};

nextset [s] = {@};

Subsequently, the QoS value of the link connected between the arbitrary node s and the arbitrary node j next to it is initialized. This also substitutes the QoS value by using a repetition sentence and determines whether the node to be initialized is its own node. If it is not its own node, it assigns a predetermined QoS value, and if it is its own node, it assigns a value of '0' (S2).

For reference, the algorithm is as follows.

for (j = 0; j <N; j ++) {

if (s == j) linkqos [s] [j] = ZERO else linkqos [s] [j] = getlinkqos ();

}

Subsequently, a set representing a node representing information of a node to be connected next to its own node is constructed using the above-described connection QoS value, which is also constructed by using a repetitive sentence for the next connected node.

That is, if the node to be connected next in the current node is not a node of its own, the node is additionally registered in the set indicating the next connecting node (S3).

For reference, the algorithm is as follows.

for (j = 0; j <N; j ++) {

if (linkqos [s] [j]! = ZERO && linkqos [s] [j]! = MAXQ) nextset [s] = nextset [s] ∪ j;

}

/ * Connected: Qos, not connected: MAXQ, 13), nextset [s]: neighbors of s * /

Subsequently, the start point node is assigned to the set representing the path search to start the path search, the value '0' is substituted for the total QoS value at this time, and when the path search is started, the initialized start point corresponds to the source node. Check whether or not (S4).

If the check result corresponds to the source node, the multipath search algorithm (MC_qospath ()) is called, and then the path search message is transmitted and waited for a signal that the path was found for two times the required delay time.

When the path discovery message REPLY is received, the highest overlapping level W is selected to notify that the path has been found, and if not, the path discovery message REPLY is notified (S5).

For reference, when the overlapping degree W is performed to a plurality of destinations, the overlapping degree W is set as the overlapping degree when going to the same node up to the middle.

That is, it refers to the node before the node where the destination node is divided by going through the same node in the same source node.

Meanwhile, if the starting point does not correspond to the source node, it belongs to the intermediate node, and thus, the corresponding path search algorithm is called according to the next message type received. At this time, the path search algorithm includes the sequential path search algorithm (MC_qospath ()), and when the current intermediate node is the last destination node, the path to the source node is transmitted to the source node without searching the path anymore. There is an algorithm MC_back_qospath () (S6).

For reference, the algorithm is as follows.

route.set = {ss};

route.tqos = ZERO;

if (ss is the source node) {

MC_qospath (ss, route, req));

while (delay 〈2 * requested_QoS_delay) {

wait_message_from_socket ();

if (received the REPLY message from all destination nodes) {

select the weightiest route from source to destination nodes;

exit (success);

}

exit (fail);

}

}

else {

while (1) {

wait_message_from_socket ();

if (message_type == REQ)

MC_qospath ();

else if (message_type == REP)

MC_back_qospath ();

}

}

FIG. 2 is a flowchart illustrating a path search process (MC_qospath ()) performed when a starting point corresponds to a source node in FIG. 1, and calculates a total QoS value up to a node searched for a path (T1).

For reference, the algorithm and variable definitions are as follows.

Qospath (int s, Route route)

{

Qos tqos;

int i;

tqos = F (route.tqos, linkqos [E (route.set), s]);

Subsequently, it is determined whether the node inputted for the path search belongs to one of the plurality of target nodes, and if the target node is satisfied, it is determined whether or not the calculated total QoS value satisfies the required reference QoS value. It notifies the source node of the path discovery and sends a REPLAY message to the immediately preceding node to find a path with high overlap among the paths generated in the middle, and sets the overlap value W to 1 (T2). .

On the other hand, if the determination result does not belong to the destination node, the total QoS that has reached the current node through a path different from the calculated total QoS value and has been calculated (because there are several paths to the node up to now) If the previously calculated value is smaller than the value, it stops the path search currently being searched and waits for another path search call (T3).

Subsequently, if the total QoS value of the currently arrived route is smaller as a result of the comparison in the third step (T3), since the route becomes the shortest distance than the last route, the route search information should be modified to this route. Check whether the last node of the path is the same as its node (T4).

If it is determined that the node is not its own node, the node is additionally registered in the connection information and the next process (T6) is performed (T5).

If it is determined that the node is its own node, the process (T5) is omitted and the total QoS value for the path up to the present is immediately calculated. When adding the connection information, the node is duplicated when the node is added.

That is, the total QoS value calculated above is substituted into the total QoS value representing the path, and the route path is substituted into the set representing the node so far (T6).

For reference, the algorithm is as follows.

if (s is in a Destination set) {

if (C (tqos) 〈C (req)) {

R_node = E (route.set);

route.set = route.set-R_node;

send (REP, R_node, route.set, W = 1);

};

};

else if (C (tqos) <C (node [s] .tqos)) / * C: Cost function for a given Qos * /

{

if (E (route.set)! = s) route.set = route.set ∪ s;

route.tqos = tqos;

node [s] .set = route.set;

}

When the process is completed, since the path search for the called node is completed, a message (i, Qospath (i, route)) is sent for path search for each neighboring node. This message is sent to part S5 which receives the call of FIG. When the call is performed according to the message, the above processes (T1 to T6) are repeated.

In particular, when the path search algorithm is repeatedly executed by the call, the path search is started after subtracting the performed route set from the set representing the connection to the next node. This is to avoid repeating the path once it has passed (T7).

For reference, the algorithm is as follows.

for (each i ∈ (nextset [s] route.set))

send (i, route);

FIG. 3 is a flowchart illustrating a path search process (MC_back_qospath ()) performed when the starting point does not correspond to the source node in FIG. 1, and determines whether there is a path having a higher degree of overlap than the degree of overlap stored in the current node. If present, the current overlapping degree value is changed to the value of the newly found overlapping path, and the overlapping degree is increased by one (T11).

Then, the current node is removed from the path so far and a path discovery message (REPLY message) is transmitted to the immediately preceding node (T12).

For reference, the algorithm is as follows.

if (node.W <W) {/ * find a path with a nesting degree higher than the nesting degree stored at the current node * /

node.W = W;

W ++;

R_node = E (route.set);

route.set = route.set-R_node;

send (REP, R_node, route, W);

}

FIG. 4 is a flowchart illustrating a disjointed multi-path search algorithm supporting QoS according to another embodiment of the present invention. The algorithm is disclosed by the applicant (Domestic Application No. 99-23623 (99.6.23). In the path search method that supports quality of service (QoS), in the algorithm that selects all paths that meet the required QoS among the paths arriving from a single source node to a single destination node, This algorithm avoids intermediate path duplication between selected multipaths.

Applying this method, there is no intermediate path overlapping each other between the various paths, there is an advantage that can prevent the situation that the multiple paths having the node is unavailable due to the failure of one intermediate node.

The process is as follows.

Initialize the total QoS value for any node s with a value of 'MAXQ', meaning that this node is a node that is not passed, a value with information about that node, and a value for the next node to be connected to this node. Initializes to an empty set (U1).

For reference, the algorithm and variable definitions are as follows.

typedef struct {Qos tqos; Oset set; } Route;

struct {Qos tqos; Oset set;} node [s];

Oset nextset [s];

Qos linkqos [s] [N];

Route route;

node [s] .tqos = MAXQ;

node [s] .set = {@};

nextset [s] = {@};

Subsequently, the QoS value of the link connected between the arbitrary node s and the arbitrary node j next to it is initialized. This also substitutes the QoS value by using a repetition sentence and determines whether the node to be initialized is its own node. If it is not its own node, it assigns a predetermined QoS value, and if it is its own node, it assigns a value of '0' (U2).

For reference, the algorithm is as follows.

for (j = 0; j <N; j ++) {

if (s == j) linkqos [s] [j] = ZERO else linkqos [s] [j] = getlinkqos ();

}

Subsequently, a set representing a node representing information of a node to be connected next to its own node is constructed using the above-described connection QoS value, which is also constructed by using a repetitive sentence for the next connected node.

That is, if the node to be connected next in the current node is not a node of its own, the node is additionally registered in the set indicating the next connecting node (U3).

For reference, the algorithm is as follows.

for (j = 0; j <N; j ++) {

if (linkqos [s] [j]! = ZERO && linkqos [s] [j]! = MAXQ) nextset [s] = nextset [s] ∪ j;

}

/ * Connected: Qos, not connected: MAXQ, 13), nextset [s]: neighbors of s * /

As described above, when all the initialization is completed, the path search algorithm is started. If the current node corresponds to the source node, the path search algorithm is determined. If the source node is the source node, the route set is initialized to the current node. Then call to perform the route search algorithm (QoSpath (s, route, req)).

If it is determined that the source node is not a source node as a result of the determination, the node, which is a value transmitted as a result of the route search process and receives the route value, is received in another node to complete the route search of the next node. Waiting to be)) When the result is received, the route search algorithm is called (QoSpath (s, route, req)) (U4, U5).

For reference, the algorithm is as follows.

if (s is the source node)

{route.set = {s};

route.tqos = ZERO;

}

else

{

receive (s, Qospath (s, route, req));

call (Qospath (s, route, req));

}

FIG. 5 is a flowchart illustrating a detailed process of the path searching process of FIG. 4, and calculates a total QoS value up to the node searched for the path (U51).

For reference, the algorithm and variable definitions are as follows.

tqos = F (route.tqos, linkqos [E (route.set), s]);

If the total QoS value to the current node searched for the path has been calculated, there are several paths that reach the current node through the path different from the total QoS value at this time. If the previously calculated value is smaller than the total QoS value, the path search that is currently being searched is stopped and another path search call is waited for (U52).

As a result of the comparison, if the total QoS value of the currently arrived path is smaller, this path is the shortest distance from the last path, and thus the path search information should be modified to this path. It is determined whether the intermediate node of the paths and the intermediate node of the previous paths overlap.

If it is determined that the duplicate node exists, the current node does not need to be added as a new path and thus stops the path search algorithm (U53).

On the other hand, if there is no duplicate node as a result of the determination, it is checked whether or not the last node of the path arriving in the current path is the same as its own node. (54).

If it is determined that the node is its own node, the process (U55) is omitted and the total QoS value for the path up to the present is immediately calculated. When adding the connection information, the node is duplicated when the node is added.

That is, the total QoS value calculated above is substituted into the total QoS value representing the path, and the value found at the appropriate position is input to the found path according to priority in the existing path set, and then the sequence is adjusted (U56). .

Subsequently, the newly discovered intermediate node information is added to all intermediate node sets passed through so far, and new path information is added to the multipath set of nodes (U57).

When the process is completed, since the path search for the called node is completed, a path search information message (i, route, req) is sent to a node not included in the intermediate path among the adjacent nodes of the node. This message is sent to the part that receives the call during the initialization process. When the call is performed according to the message, the above processes (U51 to U57) are repeated.

In particular, when the path search algorithm is repeatedly executed by the call, the path search is started after subtracting the performed route set from the set representing the connection to the next node. This is to avoid repeating the path once it has passed (U58).

For reference, the algorithm is as follows.

if (C (tqos) 〈C (node [s] .tqos [p])) {

0 = <p <m / * If m multipaths are found, p indicates the order in which m paths are listed in good order * /

if ((route.set-node [s] .sset) == (route.set-ss)) {

if (E (route.set)! = s) route.set = route.set U s;

route.tqos = tqos;

for (k = p; k <m-1; k ++) {

node [s] .tqos [k + 1] = node [s] .tqos [k];

node [s] .set [k + 1] = node [s] .set [k];

}

node [s] .sset = node [s] .sset U route.set;

node [s] .tqos [p] = tqos;

node [s] .set [p] = route.set;

for (each i ?? (nextset [s]-route.set))

send (i, route, req);

}

}

In the meantime, the path search process is described when the network configuration corresponds to the first level, and the process of performing the path search when the network configuration is the two level hierarchical network will be described below.

If the network configuration is two levels, the address of the two-level hierarchical network is first defined as a pair of integers [n, m].

Here n denotes a higher address and m denotes a lower address.

A set of nodes having the same higher address is called a region. In particular, in a hierarchical network, an address having a different upper address or a connection address list divided into addresses may not have the same upper address.

In other words, a connection outside a certain area cannot go through the same area again.

For example, routing address lists such as [1,2] [1,4] [1,3] [3,1] [3,2] [3,3] [2,1] [2,2] are hierarchical. It is suitable as an address list for structural routing, but with [1,2] [1,4] [1,3] [3,1] [2,1] [3,2] [3,3] [2,2] The same address list is not suitable for hierarchical routing because the address groups have [3 ,.] separated by [2,1] with the same upper address (different higher address).

In other words, this connection leaves node 3 through node [3,1], passes through node [2, 1], and enters node 3 again through node [3, 2]. That's because you can't go into the same area.

Hereinafter, a path search method at two levels will be described.

FIG. 6 is a flowchart illustrating a multipath discovery (MCHDGSPA) process supporting QoS in a two-level network according to a second embodiment of the present invention, which is almost similar to the above-described MCDGSPA, and the other part is initialized. You need to initialize both levels.

A total QoS value for any node p belonging to any layer i is initialized to a value of 'MAXQ' which means that this node is a node that is not passed, and a set having information about any node belonging to any layer. In this case, the set indicating the node belonging to the next layer to be connected with this node is initialized to an empty set (V1).

For reference, the algorithm and variable definitions are as follows.

struct {Qos tqos; Oset set;} node [H, N]; / * Hierarchical address * /

Qos linkqos [[H, N], [H, N]];

Oset nextset [H, N];

struct Route {Qos tqos; Oset set; } route;

struct Node {int s; int p;};

node [i, p] .tqos = MAXQ;

node [i, p] .set = {@};

nextset [i, p] = {@};

Subsequently, the QoS value of the link connected between any node p belonging to the arbitrary layer i and any node q belonging to any layer j next to it is initialized, which is also assigned a QoS value using a repetitive statement. However, it determines whether the node of the layer to be initialized is its own node, and if it is not its own node, substitutes a preset QoS value, and if it is its own node, substitutes a value of '0' (V2).

For reference, the algorithm is as follows.

for (i = 0; i <H; i ++) for (j = 0; j <H; j ++) {

for (p = 0; p <N; p ++) for (q = 0; q <N; q ++) {

if ((i, p) == (j, q)) linkqos [[i, p], [j, q]] = ZERO;

else linkqos [[i, p], [j, q]] = getlinkqos ();

}

Subsequently, a set representing node information of a layer to be connected next to its own node is constructed using the above-described connection QoS value, which is also constructed by using a repetition sentence for nodes of a layer connected to the next.

That is, if the node of the layer to be connected next in the node of the current layer is not a node of its own, the node is additionally registered in the set indicating the next connecting node (V3).

For reference, the algorithm is as follows.

for (i = 0; i <H; i ++) for (j = 0; j <H; j ++) {

for (p = 0; p <N; p ++) for (q = 0; q <N; q ++) {

if (linkqos [[i, p], [j, q]]! = ZERO & & linkqos [[i, p], [j, q]]! = MAXQ)

nextset [i, p] = nextset [i, p] ∪ (j, q);

}

}

/ * Connected: Qos value, not connected: MAXQ * /

Subsequently, the start point node is assigned to the set representing the path search to start the path search, the value '0' is substituted for the total QoS value at this time, and when the path search is started, the initialized start point corresponds to the source node. Check whether it is (V4).

If the check result corresponds to the source node, the multipath search algorithm (MC_qospath ()) is called, and then the path search message is transmitted and waited for a signal that the path was found for two times the required delay time.

When the path discovery message REPLY is received, the overlapping degree W selects the highest path and notifies that the path has been found. If not, the path discovery message REPLY is notified (V5).

For reference, when the overlapping degree W is performed to a plurality of destinations, the overlapping degree W is set as the overlapping degree when going to the same node up to the middle.

That is, it refers to the node before the node where the destination node is divided by going through the same node in the same source node.

Meanwhile, if the starting point does not correspond to the source node, it belongs to the intermediate node, and thus, the corresponding path search algorithm is called according to the next message type received. At this time, the path search algorithm includes the sequential path search algorithm (MC_qospath ()), and when the current intermediate node is the last destination node, the path to the source node is transmitted to the source node without searching the path anymore. There is an algorithm (MC_back_qospath ()) (V6).

For reference, the algorithm is as follows.

route.set = {ss};

route.tqos = ZERO;

if (ss is the source node) {

MC_qospath (ss, route, req));

while (delay 〈2 * requested_QoS_delay) {

wait_message_from_socket ();

if (received the REPLY message from all destination nodes) {

select the weightiest route from source to destination nodes;

exit (success);

}

exit (fail);

}

}

else {

while (1) {

wait_message_from_socket ();

if (message_type == REQ)

MC_qospath ();

else if (message_type == REP)

MC_back_qospath ();

}

}

Subsequently, a path search process performed when the starting point corresponds to the source node and a path search process performed when the starting point does not correspond to the source node in FIG. 6 are the same as the processes shown in FIGS. The only difference is that the path discovery message is transmitted in step T7 of FIG. 2 to repeat the path search.

Since this is a path search on a two-level network, the path in the layer that has been passed once is not repeated again. For example,

{(S, p), (w, r), (f, g), (h, k)}-{(c, p), (w, r), (f, v), (h, r)} = {(s, p), (h, k)}. This means that the hierarchical layer does not go through the hierarchical path, and the w and f hierarchies have already gone through, so it does not go through again. .

FIG. 7 is a flowchart illustrating a disjointed multi-path discovery process supporting QoS in a two-level network according to a third embodiment of the present invention, which is an initialization part V1 to V3 process of FIG. Until the same and only the next part is different, the same parts will be omitted.

When all the initialization is completed through the steps V1 to V3, a path search algorithm is started. If the node of the current layer corresponds to a source node, the path search algorithm is determined. If the source node is a source node, the path search algorithm is initialized to the current node. The corresponding QoS value is also initialized to '0' and then called to perform the path search algorithm (QoSpath (s, route)).

If it is determined that the source node is not the source node, the node completes the route search and receives the node, which is the value transmitted as a result of the route search process, and the route value at this time for receiving the route of the next node (receive (s, route)). Wait to be. When the result is received, the route search algorithm is called (QoSpath (s, route)) (V7, V8).

For reference, the algorithm is as follows.

if (s is the source node)

{route.set = {s};

route.tqos = ZERO;

}

else

{

receive (s, Qospath (s, route));

call (Qospath (s, route));

}

Subsequently, the detailed process of the route search process in FIG. 7 is the same as the process shown in FIG. 5, and thus description thereof is omitted. The only difference is that the route search message is transmitted in process T7 of FIG. 2 to repeat the route search. There is a difference in the process.

Since this is a path search on a two-level network, a path in a layer that has been once passed is not repeated again. This is the same as an example described in the path search process of FIG.

As described above, the path searching method provided by the present invention has been described. In addition, the comparison object in the cost function C is present between a delay value for receiving a reception value for message transmission and a destination node in the source node. In the case of the number of hops, the value generated in the current path may be smaller than the value generated in the existing path. In the case of comparing bandwidth values, the value generated in the current path may be larger than the value generated in the existing path.

These conditions can be selectively used according to the users using the respective path searching methods. Also, the MCHDGSPA and DMHDGSPA algorithms in the second level layer can be used as an example for the case where the network is in the second level. For example, the level of the network can be configured as many levels as possible, and accordingly, the present invention can also be applied to multiple levels because the set is increased according to the level.

As described in detail above, the present invention provides a shortest path in a single-layer network in order to provide QoS-based routing of routing required for transmitting multimedia information on various information communication networks. In the case where there are multiple nodes (MCDGSPA), and in the case of setting up multiple paths that satisfy the baseline QoS that may occur for a single purpose in a single source node (QRDGSPA), the same node may be included from the middle node. If there is a path, one of the nodes provides methods to delete (DMDGSPA) and provides the shortest path search applied to the multilevel layer. By providing respective algorithms (MCHDGSPA, DMHDGSPA), each of the above path search algorithms Standing has the advantage of allowing users to selectively use the algorithm that best matches your system environment.

In addition, preferred embodiments of the present invention are disclosed for the purpose of illustration, those skilled in the art will be able to various modifications, changes, additions, etc. within the spirit and scope of the present invention, such modifications and modifications belong to the following claims You will have to look.

Claims (28)

In the path search method on an information communication network, A first process of initializing a set and quality of service (QoS) values indicating connection information of nodes in all paths existing in the network to search the shortest paths of multiple paths occurring from a single source node to multiple destination nodes. and; A second process of calling a route search means by designating a starting point for the route search and receiving a result message from the route search means to distribute the route search when the initialization is completed; And performing a path search according to the call, and then transmitting a result message to the call target. The method of claim 1, The first step of initializing total QoS values accumulated in any node along the shortest route from the source from node 0 to node N; A second step of initializing a value having information about the arbitrary node and a value indicating a next node to be connected with the node; A third step of initializing a QoS value of a link connected between the any node and any node next to it; And a fourth step of constructing a set representing information of a node to be connected next to its own node using the QoS value initialized in the third step. The method of claim 2, In the fourth step, when a set representing information about a node to be connected next to the node is constructed, the next node is connected if the node to be connected next to the current node is not a node of its own. Path searching method supporting quality of service (QoS), characterized in that the registration in addition to. The method of claim 1, In the second process, when the initialization is completed in the first process, a node to start the route search is assigned to the set representing the route search to perform the route search, and the value 0 is substituted in the total QoS value. A first step of determining whether the initialized starting point corresponds to a source node; A second step of transmitting a multi-path search request message for starting a path search from the source node if the source node corresponds to the source node and waiting for a path discovery signal to be received for a specific time of the required delay time; And a third process of invoking the multipath search means and the overlapping comparison path search means according to the received message type if the starting point does not correspond to the source node. How to navigate the route. The method of claim 4, wherein The path search step of the multipath search means in the third step includes: a first step of calculating a total QoS value up to the node searched for the path; Determining whether the input node belongs to one of a plurality of target nodes, and if the target node determines whether the total QoS value satisfies the required reference QoS value; If the value does not belong to the target node, the total QoS value is different from the total QoS value, which is reached to the current node through a different path. A third process of waiting for; A fourth step of determining whether or not the last node of the route arriving at the current route is the same as its own node in order to modify the route search information if the total QoS value of the currently arrived route is smaller as a result of the comparison; A fifth step of additionally registering the node in the connection information if the node is not its own node and omitting the additional registration if the node is its own node; A sixth step of substituting the total QoS value calculated in the first step into the total route QoS value for the path up to now and registering the route path in a set representing the node up to now; After each process is completed, the route set is subtracted from the set representing the connection to the next node for the route search to the next node of the called node. And a seventh process of transmitting to a place, wherein the route searching method supports quality of service. The method of claim 5, If the reference QoS value is satisfied in the second process, a path discovery message is notified to the source node, and a path discovery message is transmitted to the immediately preceding node in order to find a path having a high overlapping rate among intermediate paths, and the overlapping value is transmitted. Path discovery method supporting quality of service (QoS), characterized in that the change to 1. The method of claim 4, wherein When receiving the path discovery message in the second process, the path having the highest overlap is selected and notified that the path is found, and if not, the path discovery message is notified. How to navigate the route. The method of claim 4, wherein In the third process, the process of searching for the overlapping degree comparison path determines whether there is a path having a higher degree of overlap than the degree of overlapping stored in the current node. A first process of changing to and increasing the degree of overlap by one; And a second process of removing the current node from the path up to the present and transmitting a path discovery message to the immediately preceding node after the above process. In the path search method on an information communication network, Aggregation and Quality of Service (QoS) indicating the connection information of nodes in all paths existing in the network to search all paths satisfying the baseline QoS value among the multiple paths that can occur from a single source node to a single destination node. A first step of initializing a value; A second step of calling the route search means by specifying a starting point to search for the route and a reference QoS value to distribute the route search when the initialization is completed, and receiving a result message from the route search means; And performing a path search according to the call, deleting the remaining paths except one of the paths in which duplicate nodes exist among the found plurality of paths, and then transmitting a result message to the call target. Path discovery method that supports quality of service (QoS). The method of claim 9, The first step of initializing total QoS values accumulated in any node along the shortest route from the source from node 0 to node N; A second step of initializing a value having information about the arbitrary node and a value indicating a next node to be connected with the node; A third step of initializing a QoS value of a link connected between the any node and any node next to it; And a fourth step of constructing a set representing information of a node to be connected next to its own node using the QoS value initialized in the third step. The method of claim 10, In the fourth step, when a set representing information about a node to be connected next to the node is constructed, the next node is connected if the node to be connected next to the current node is not a node of its own. Path searching method supporting quality of service (QoS), characterized in that the registration in addition to. The method of claim 9, The second process may include: a first process of determining whether a current node corresponds to a source node when initialization is completed in the first process; A second step of initializing a set representing a path as a current node if the determination result is a source node, initializing a QoS value of the set to '0', and then calling a path searching unit; And a third process of calling a path search means by using a set and a value indicating a node which is a result value transmitted at the time of completing a path search for a previous node and a path at this time if it is not the source node as a result of the determination. Path discovery method that supports quality of service (QoS). The method of claim 9, The path search process performed in the third process may include a first process of calculating a total QoS value of a node searched for a path to date; When the total QoS value is calculated, if the calculated value is smaller than the total QoS value previously reached and reached to the current node through a path different from the total QoS value at this time, the search for the current path is stopped and the other path is stopped. A second process of waiting for a search call; A third step of determining whether the intermediate node of the newly discovered paths and the intermediate node of the previous paths overlap if the total QoS value of the currently arrived path is smaller as a result of the comparison; If there is no duplicate node as a result of the determination, determining whether the last node of the route arriving at the current route is the same as its own node; A fourth step of additionally registering the node in the connection information if the node is not its own node and omitting the additional registration if the node is its own node; A fifth step of substituting the total QoS value calculated in the first step into the total route QoS value for the current path after the step 4 and inputting the found path to a position according to priority in the existing path set; and; A sixth step of adding the newly found intermediate node information to all the intermediate node sets passed through to date and adding the new path information to the multipath set of nodes after the fifth step; Quality of Service (QoS) characterized in that it comprises a seventh process of transmitting the resulting message to the place where the path search means is called for the path search for the next node of the called node when each process is completed. Path navigation method that supports. The method of claim 13, If there is a duplicate node as a result of the determination in the third process, the current node does not need to add a new path, and thus the path search process is interrupted. In the path search method on an information communication network, Represents connection information of nodes in layers in all paths existing on the multi-level network to search the shortest paths of the multiple paths occurring from a single source node to a plurality of destination nodes on a network in which the layer of the network is multilevel. A first step of initializing an aggregate and a quality of service (QoS) value; A second process of calling a route search means by designating a starting point for the route search and receiving a result message from the route search means to distribute the route search when the initialization is completed; When the route search is performed according to the call, the layer of the layer that has once passed through is not re-routed, and when the route search is performed, a third process of transmitting a result message to the call target is provided. Supported route navigation methods. The method of claim 15, The first step of initializing total QoS values accumulated in nodes in an arbitrary layer along the shortest route from the source from node 0 to node N; A second step of initializing a value having information about a node in the arbitrary hierarchy and a value indicating a next node to be connected with the node in the hierarchy; A third step of initializing a QoS value of a link connected between the node in any layer and any node next to it; And a fourth step of constructing a set representing information of a node to be connected next to its own node using the QoS value initialized in the third step. The method of claim 16, In the fourth step, when a set representing information about a node to be connected after the node is constructed, if the node in the layer to be connected next is a node in which the path is connected without being a node of the node in the current layer, the next connection node. Path searching method supporting quality of service (QoS), characterized in that the additional registration to the set indicating. The method of claim 15, In the second process, when the initialization is completed in the first process, a node to start the route search is assigned to the set representing the route search to perform the route search, and the value 0 is substituted in the total QoS value. A first step of determining whether the initialized starting point corresponds to a source node; A second step of transmitting a multi-path search request message for starting a path search from the source node if the source node corresponds to the source node and waiting for a path discovery signal to be received for a specific time of the required delay time; And a third process of invoking the multipath search means and the overlapping comparison path search means according to the received message type if the starting point does not correspond to the source node. How to navigate the route. The method of claim 18, The path search step of the multipath search means in the third step includes: a first step of calculating a total QoS value up to nodes in the layer searched for to date; Determining whether an input node in the hierarchy belongs to one of a plurality of target nodes, and if the target node determines whether the total QoS value satisfies a required reference QoS value; If the value does not belong to the target node, the total QoS value is different from the total QoS value, which is reached to the current node through a different path. A third process of waiting for; A fourth step of determining whether or not the last node of the route arriving at the current route is the same as its own node in order to modify the route search information if the total QoS value of the currently arrived route is smaller as a result of the comparison; A fifth step of additionally registering the node in the connection information if the node is not its own node and omitting the additional registration if the node is its own node; A sixth step of substituting the total QoS value calculated in the first step into the total route QoS value for the path up to now and registering the route path in a set representing the node up to now; After each process is completed, the route set means is subtracted from the set representing the connection to the next node for the route search for the next node of the node in the called hierarchy. And a seventh process of transmitting to the called place. The method of claim 19, If the reference QoS value is satisfied in the second process, a path discovery message is notified to the source node, and a path discovery message is transmitted to the immediately preceding node in order to find a path having a high overlapping rate among intermediate paths, and the overlapping value is transmitted. Path discovery method supporting quality of service (QoS), characterized in that the change to 1. The method of claim 18, When receiving the path discovery message in the second process, the path having the highest overlap is selected and notified that the path is found, and if not, the path discovery message is notified. How to navigate the route. The method of claim 18, In the third process, the process of searching for the overlapping degree comparison path determines whether there is a path having a higher degree of overlap than the degree of overlapping stored in the current node. A first process of changing to and increasing the degree of overlap by one; And a second process of removing the current node from the path up to the present and transmitting a path discovery message to the immediately preceding node after the above process. In the path search method on an information communication network, Nodes on all paths existing on the multi-level network to search all paths satisfying the reference QoS value among the multiple paths that can occur from a single source node to a single destination node on a network where the layer of the network is multilevel. A first step of initializing a set indicating a connection information of the service and a quality of service (QoS) value; A second step of calling the route search means by specifying a starting point to search for the route and a reference QoS value to distribute the route search when the initialization is completed, and receiving a result message from the route search means; When performing a path search according to the call, the area of a layer that has been one-time does not go through again, and deletes the remaining paths except one of the paths where duplicate nodes exist among a plurality of searched paths, and then outputs the result message. And a third process of transmitting the data to a path search method supporting quality of service (QoS). The method of claim 23, wherein The first step of initializing total QoS values accumulated in nodes in an arbitrary layer along the shortest route from the source from node 0 to node N; A second step of initializing a value having information about a node in the arbitrary hierarchy and a value indicating a next node to be connected with the node in the hierarchy; A third step of initializing a QoS value of a link connected between the node in any layer and any node next to it; And a fourth step of constructing a set representing information of a node to be connected next to its own node using the QoS value initialized in the third step. The method of claim 24, In the fourth step, when a set representing information about a node to be connected after the node is constructed, if the node in the layer to be connected next is a node in which the path is connected without being a node of the node in the current layer, the next connection node. Path searching method supporting quality of service (QoS), characterized in that the additional registration to the set indicating. The method of claim 23, wherein The second process includes a first process of determining whether a node in a current layer corresponds to a source node when initialization is completed in the first process; A second step of initializing a set indicating a path to a current node if it is a source node, initializing a QoS value of the set to '0', and then calling a path searching unit; And a third process of calling a path search means by using a set and a value indicating a node which is a result value transmitted at the time of completing a path search for a previous node and a path at this time if it is not the source node as a result of the determination. Path discovery method that supports quality of service (QoS). The method of claim 23, wherein The path search process performed in the third process may include a first process of calculating a total QoS value of a node searched for a path to date; When the total QoS value is calculated, if the calculated value is smaller than the total QoS value previously reached and reached to the current node through a path different from the total QoS value at this time, the search for the current path is stopped and the other path is stopped. A second process of waiting for a search call; A third step of determining whether the intermediate node of the newly discovered paths and the intermediate node of the previous paths overlap if the total QoS value of the currently arrived path is smaller as a result of the comparison; If there is no duplicate node as a result of the determination, determining whether the last node of the route arriving at the current route is the same as its own node; A fourth step of additionally registering the node in the connection information if the node is not its own node and omitting the additional registration if the node is its own node; A fifth step of substituting the total QoS value calculated in the first step into the total route QoS value for the current path after the step 4 and inputting the found path to a position according to priority in the existing path set; and; A sixth step of adding the newly found intermediate node information to all the intermediate node sets passed through to date and adding the new path information to the multipath set of nodes after the fifth step; Quality of Service (QoS) characterized in that it comprises a seventh process of transmitting the resulting message to the place where the path search means is called for the path search for the next node of the called node when each process is completed. Path navigation method that supports. The method of claim 27, If there is a duplicate node as a result of the determination in the third process, the current node does not need to add a new path, and thus the path search process is interrupted.