JP2003218915A - Traffic load distribution system and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To distribute load on traffic by distributing data flows on a path on which the traffic is concentrated to a route group bypassed for each packet type. <P>SOLUTION: A management terminal 200 sets packet type information defined by combinations of header information of packets, defines inter-router access information for each packet as to routers configuring a network, and transmits the packet type information and the inter-router access information to routers 30 to 90. Each router creates a virtual routing table for each packet type on the basis of the inter-router access information. Upon receipt of a packet by each router relaying the packet sent from a terminal 10 to a terminal 20, each router searches whether or not the header information of the packet is registered in the packet type information. When the header information is registered in the packet type information, each router uses the virtual routing table corresponding to the packet type to route the packet. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a traffic load balancing method and method for balancing the traffic load of a network, and more particularly to dynamically collecting network topology information by using a routing protocol and TOS (Type
The present invention relates to a traffic load balancing method and method that guarantees the CoS (Class of Service) required for traffic by selecting the shortest route in consideration of the (of service) and using traffic engineering together.

[0002]

2. Description of the Related Art In a connectionless packet network (IP (Internet Protocol) network), a conventional OSPF (IETF: RFC1247, Open Shortest Path Fir) is used.
In the link state protocol represented by st), the following method is used to select the optimum route for TOS for each data flow. The routers that compose the network calculate the cost of delay, throughput, and reliability for each TOS for all output ports and notify the cost information to the entire network. The router belonging to the network calculates the shortest route based on the notified cost for each TOS and generates the routing table. As a result, it is possible to dynamically select the shortest path suitable for the TOS of the data flow based on the cost change due to congestion or network topology change in the network.

Diffserv (IETF: RFC2474, RFC2475, Differe) is used as a means for guaranteeing QoS (Quality of Service) for packets passing through the shortest route obtained by the above-mentioned routing protocol.
ntiated Services) and RSVP (IETF: RFC2205-2209, Resourc
e ReSerVation Protocol).

In Diffserv, as an agreement between the network administrator and the user in advance, the processing method for the received packet in the router in the Diffserv domain and the I
The relationship of DSCP (DiffServ Code Point) in the P header is defined. As a result, differentiated services are realized for each DSCP. As service classes, there are EF (Expedited Forwarding) for guaranteeing bandwidth and delay and AF (Assured Forwarding) for realizing relative priority processing. Also, DS
As CP, the Type of Service field in the IPv4 header is used in IPv4, and the Priority field in the IPv6 header is used in IPv6. Fig. 15 shows the IPv4 header format (R
FC791) and the IPv6 header format (RFC188
3) is shown.

[0005] RSVP also reserves bandwidth for each data flow in the router through which it passes.
It is a means to guarantee S. When the transmitting terminal transmits a data flow, the routing protocol operating in the network determines the route through which the data flow passes. In RSVP, the IP address information of the router at the previous stage is stored in each router on this route.
As a result, the route for reserving the band is determined. When the route through which the data flow passes is changed by the routing protocol, each router on the network updates the accumulated IP address information of the preceding router. The receiving side terminal requests the bandwidth required for the data flow from the router on the route determined above. This makes it possible to guarantee CoS for a specific data flow.

[0006]

However, the above-mentioned conventional techniques have the following problems.

The first problem is that in the conventional routing protocol, it is possible to specify a route group that is permitted as a passage route and a route group that is not permitted in all routes that can be assumed as the passage route of the packet. I could not do it. Further, in the conventional static routing, only fixed routes registered in advance can be defined, so that it is not possible to dynamically respond to changes in network conditions. Therefore, it has been impossible to perform flexible route control such as grouping some routes that are candidates for passage and automatically selecting an optimal route among them.

[0008] The second problem is that all routers on the network are required to perform optimal route routing for each TOS using a link state protocol such as OSPF.
It is necessary to calculate the cost for each TOS at all output ports that are always implemented and notify the entire network.
In addition, all routers on the network must constantly calculate the optimum route for each TOS from the received cost information and create a routing table for each TOS. Since these processes increase the load of the processes required for the entire network, TOS routing has not been used very often.

The third problem is that when TOS routing is performed in a link state protocol such as OSPF,
Since the routing is only for each TOS, the same route is applied to all packets with the same TOS. As a result, due to features such as packet length that are not reflected in TOS,
It was not possible to control the distribution of packet routes to distribute the load within the network.

The fourth problem is that, in traffic engineering such as DiffServ and RSVP, existing routing protocols such as OSPF are used as a route selection means, so that the problems 1 to 3 pointed out above are solved. I couldn't do it.

A first object of the present invention is, in order to solve the above-mentioned first problem, according to a specific packet type, as a route through which the packet type passes, a permitted route group and a non-permitted route group. By providing a function of designating a group, a traffic load balancing method and method for performing flexible route control suitable for network characteristics are provided.

In order to solve the above-mentioned second problem, a second object is to previously define a route group for bypassing a route in which traffic is concentrated for each packet type by the route control method according to the present invention. By
Providing a traffic load balancing method and method that performs traffic load balancing that simplifies the processing required for the entire network by omitting the procedure that constantly calculates the optimal route based on the cost of each TOS, which has been a bottleneck in the past That is.

As a third object, in order to solve the above-mentioned third problem, by handling arbitrary IP packet header information in addition to the TOS of the IP packet as a parameter defining the packet type, it is possible to achieve flexibility. A traffic load balancing method and method for selecting a packet type.

As a fourth object, in order to solve the above-mentioned fourth problem, by adding the functions described for the first to third objects to the conventional routine protocol, the conventional traffic engineering is performed. Using the function of the present invention,
It is an object of the present invention to provide a traffic load balancing method and method for realizing QoS control.

[0015]

A traffic load balancing system according to the first invention of the present application is a connectionless packet network (IP network) that dynamically collects network topology information using a routing protocol to select the shortest route. ), The traffic load balancing method according to claim 1, further comprising: means for defining a plurality of virtual networks for the network; and means for distributing traffic to the virtual network.

The traffic load balancing method of the second invention of the present application is a traffic load balancing method in a connectionless packet network (IP network) that dynamically collects network topology information using a routing protocol and selects the shortest route. The network comprises a terminal for transmitting and receiving packets, a router for relaying packets between the terminals, and a management terminal, wherein the management terminal sets means for setting packet type information for identifying packets, and the packet There is provided means for setting valid / invalid information of each inter-router connection for constructing a virtual network for each type, and means for transmitting the packet type information and the inter-router connection information for each packet type to the router. However, the router is a router that supports the existing network. Network topology table and a real routing table based on the real network topology table; a means for receiving the packet type information and inter-router connection information for each packet type from the management terminal; Means for generating a virtual network topology table for each packet type based on the network topology table and router connection information for each packet type, and creating a virtual routing table based on the virtual network topology table; Means for searching whether or not the received packet corresponds to the packet type, and if the packet type corresponds to the packet type, performs routing of the received packet using a virtual routing table corresponding to the applicable packet type It has, characterized in that.

The traffic load balancing method of the third invention of the present application is a traffic load balancing method in a connectionless packet network (IP network) that dynamically collects network topology information using a routing protocol and selects the shortest route. The network includes a terminal for transmitting and receiving packets, a router for relaying packets between the terminals, and a management terminal. The management terminal collects network topology information from the network and recognizes an existing network topology. A real network recognizing means for creating a real network topology table, a means for setting packet type information for identifying a packet, and valid / invalid information for connection between routers for constructing a virtual network for each packet type are set. You Means, a virtual network defining means for creating a virtual network topology table for each packet type based on the inter-router connection information and the actual network topology table, and the packet type information and the virtual network topology table for each packet type. Means for transmitting to the router, the router generating a real network topology table corresponding to an existing network, and creating a real routing table based on the real network topology table; and the packet type information. And means for receiving a virtual network topology table for each packet type from the management terminal, and a virtual network for each packet type based on the virtual network topology table for each packet type Means for creating a corresponding virtual routing table, and when the packet is received, it is searched whether the received packet corresponds to the packet type, and if it corresponds to the packet type, the virtual corresponding to the applicable packet type Means for routing the received packet using the routing table.

The traffic load distribution system according to the fourth invention of the present application is the traffic routing system according to the second or third invention, wherein the router is an actual routing table corresponding to the existing network when the received packet does not correspond to the packet type. It is characterized by having a means for routing a packet received by using.

The traffic load distribution system according to the fifth aspect of the present invention is characterized in that, in the second or third aspect, the packet type information comprises a combination of packet header information.

A traffic load balancing method according to a sixth aspect of the present invention is a traffic load balancing method in a connectionless packet network (IP network) which dynamically collects network topology information using a routing protocol and selects the shortest route. The router collects the network topology information from the network, creates the real network topology table corresponding to the existing network, and creates the real routing table based on the real network topology table. Set packet type information for identification, set inter-router connection information for constructing a virtual network for each packet type, and set the packet type information and the inter-router connection information for each packet type on the network. And sent to each router constituting the click, the router receives the packet type information and router connection information for each of the packet type from the management terminal,
A virtual network topology table for each packet type is generated based on the real network topology table and router-to-router connection information for each packet type, a virtual routing table corresponding to the virtual network topology table is generated, and the router When a packet is received, a search is performed to see if the packet received corresponds to the packet type. If the packet type corresponds to the packet type, the packet received using the virtual routing table corresponding to the applicable packet type It is characterized by performing routing.

A traffic load balancing method of the seventh invention of the present application is a traffic load balancing method in a connectionless packet network (IP network) which dynamically collects network topology information using a routing protocol and selects the shortest route. The router collects the network topology information from the network, creates the real network topology table corresponding to the existing network, and creates the real routing table based on the real network topology table. Collects network topology information, recognizes the existing network topology, creates an actual network topology table, sets packet type information for identifying packets, and sets the packet type information for each packet type. Setting connection information between routers for constructing a virtual network, creating a virtual network topology table for each packet type based on the connection information between routers and the actual network topology table, the packet type information and the packet type Each virtual network topology table is transmitted to each router constituting the network, the router receives the packet type information and the virtual network topology table for each packet type from the management terminal,
A virtual routing table corresponding to the virtual network for each packet type is created based on the virtual network topology table for each packet type. When the router receives a packet, the received packet corresponds to the packet type. It is characterized in that whether or not the received packet is searched and if the packet type corresponds to the packet type, the received packet is routed using the virtual routing table corresponding to the packet type.

In the traffic load balancing method of the eighth invention of the present application, in the sixth or seventh invention, when the received packet does not correspond to the packet type, the packet is received by using a routing table corresponding to an existing network. It is characterized in that the packet routing is performed.

The traffic load distribution method according to the ninth invention of the present application is characterized in that, in the sixth or seventh invention, the packet type information is a combination of packet header information.

The traffic load distribution method of the tenth invention of the present application is the method of the sixth or seventh invention, wherein the received packet corresponds to the packet type and uses the virtual routing table of the corresponding virtual network. When the destination is unreachable, the received packet is routed using the real routing table corresponding to the existing network.

[0025]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The configuration of the first embodiment of the present invention will be described in detail with reference to the drawings. FIG. 1 is a diagram showing a configuration of a first embodiment of the present invention, and represents a network topology (hereinafter, a virtual network topology, which will be described later, will be referred to as a real network topology). FIG. 2 is a diagram showing the configuration of the router. FIG. 3 is a diagram showing the configuration of the management terminal. FIG. 4 is a diagram showing an example of the packet type table.

Referring to FIG. 1, the first embodiment of the present invention is a terminal 10, a terminal 20, a router 30, a router 4.
0, router 50, router 60, router 70, router 8
0, a router 90 and a management terminal 200.

The terminal 10 is an information processing device that transmits a packet addressed to the terminal 20.

The terminal 20 is an information processing device that receives a packet from the terminal 10.

The routers 30 to 90 are devices for relaying packets addressed to the terminal 20 from the terminal 10. Router 3
0 to router 90 are router function units 1 as shown in FIG.
01, the real routing table 102, the first virtual routing table 1031, the second virtual routing table 1032, the packet type table 104, and the interface unit 105.

The router function unit 101 uses the real routing table 102 or the packet type table 104 and the virtual routing table 103 to perform packet relay processing. The router function unit 101 includes a CPU and a C
A function of creating a virtual routing table 103 based on a virtual network topology table, which is configured by a memory that stores programs executed by the PU, data, etc.
It has a function of determining the packet type of the received packet based on the packet type table 104, a function of determining a relay route of the packet based on the routing table, and a function of transmitting / receiving the packet. Further, a function of collecting connection information between routers in the network, recognizing an existing network topology, generating a real network topology table, and further creating a corresponding real routing table 102 based on the generated real network topology table. Have.

The real routing table 102 is a routing table corresponding to the real network topology and stores relay route information of packets. Each row (referred to as an entry) forming the real routing table shows information on a certain relay route.

The virtual routing table 103 is a routing table corresponding to a virtual network topology and stores packet relay route information. Each line that makes up the virtual routing table (called an entry)
Indicates information about a relay route. The virtual routing table corresponds to the virtual network topology and exists for each virtual network topology. In FIG. 2, two virtual routing tables (first virtual routing table 1031 and second virtual routing table 1032) are illustrated.

The packet type table 104 stores the packet type information transmitted from the management terminal 200, and associates the packet type with the virtual network. FIG. 4 shows an example of the packet type table. The IP header information and the virtual network number corresponding to it are registered in the packet type table. In the example of FIG. 4, Sour
The IP packet having the ce Address “10.40.29.109” is associated with the virtual network 1.

The interface section 105 connects the router to the network via a cable. There are a plurality of interface units 105 according to the number of ports.

The routers 30 to 90 generally operate as follows. When the interface unit 105 receives the packet, the router function unit 101 determines the packet type, and determines the corresponding virtual network from the packet type. The destination network is determined from the destination address of the received packet, and the virtual network routing table is searched for in the destination network. If there is no matching entry, the forwarding process is performed using the routing table of the real network, and if there is a matching entry, the entry on the routing table of the virtual network is selected and transmitted from the interface unit 105 to the forwarding destination. To do. The management information packet issued from the management terminal 200 is routed using the actual routing table. It also receives packet type information and inter-router connection information from the management terminal 200, generates a packet type table and a virtual network topology table, and further generates a corresponding virtual routing table based on the generated virtual network topology table.

The management terminal 200 is an information processing apparatus that manages a network composed of the terminals 10 to 20 and the routers 30 to 90. As shown in FIG. 3, the management terminal 200 has a control unit 201, a packet type information setting unit 202, an inter-router connection information setting unit 203, a communication unit 204, a real network recognition unit 205, and a virtual network definition unit 206.

The control means 201 controls the packet type information setting means 202, the inter-router connection information setting means 203, the communication means 204, the real network recognizing means 205 and the virtual network defining means 206.

The packet type information setting means 202 sets the packet type information for identifying the packet. Packet type information includes packet priority (TOS), data length,
It is defined by a combination of IP packet header information such as protocol type. The packet priority is IPv4.
TOS field in header or Prior in IPv6 header
Specified in the ity field. The data length is the Total Length field in the IPv4 header or Payload in the IPv6 header.
Specified in the oad Length field. The protocol type is identified by the Protocol field in the IPv4 header or the Next Header field in the IPv6 header.

The inter-router connection information setting means 203 sets valid inter-router connection information (or invalid inter-router connection information) for each packet type. The inter-router connection information indicates whether the connection between the routers in the existing network topology shown in FIG. 1 is valid or invalid as the inter-router connection when performing a route calculation of a specific packet type. Represent. The router-to-router connection information is used to generate the virtual network topology.

The communication means 204 transmits information such as the packet type information set by the packet type information setting means 202 and the inter-router connection information set by the inter-router connection information setting means 203 to the router. Further, it receives information from the terminals 10 to 20 and the routers 30 to 90.

The real network recognition means 205 dynamically collects network topology information, recognizes the existing network topology, and creates a real network topology table. The topology table includes the recognized inter-router links, cost information, and valid / invalid information of the inter-router links. The valid / invalid information of the inter-router link is a parameter peculiar to the present invention and indicates whether the inter-router link is valid or invalid.

The virtual network defining means 206 is the inter-router connection information set by the inter-router connection information setting means 203 for each packet type and the real network recognizing means 2
A virtual network topology table for each packet type is created based on the real network topology table created by 05.

The operation of the first embodiment of the present invention will be described in detail with reference to the drawings. FIG. 5 is a diagram explaining the operation of the first embodiment of the present invention. FIG. 6 is a diagram showing an example of the packet type table. FIG. 7 is a diagram showing an example of a virtual network topology. Figure 8
It is a figure which shows the flow of IP packet routing in a router. FIG. 9 is a diagram showing a changed passage route in the traffic 100. FIG. 10 is a diagram showing a changed transit route in the traffic 110. FIG. 11 is a diagram showing a traffic passage route.

First, the overall flow of operation will be described. The embodiment of the present invention operates as follows. (Step 1) The network starts up, and an actual routing table is created by a known method to perform routing. (Step 2) The management terminal 200 defines a packet type and inter-router connection information for each packet type, and notifies each router. (Step 3) Each router creates a virtual network topology table from the packet type received from the management terminal 200 and the inter-router connection information for each packet type,
Further, a corresponding virtual routing table is created. (Step 4) Each router performs routing using the real routing table and the virtual routing table.

Next, each of the above four steps will be described in detail.

First, step 1 will be described. (1-1) Referring to FIG. 5, each router dynamically collects network topology information and recognizes an existing network topology, and an actual routing table 102 based on the actual network topology is used by an algorithm of a conventional routing protocol. To create. (1-2) Using the actual routing table 102, routing is performed by the shortest route selected based on the route calculation result by the existing routing protocol (at the time when the network of FIG. Route control has not been done yet). In this network, traffic 100, traffic 110, and traffic 120 flow from the terminal 10 to the terminal 20. These traffics are groups of IP packet datagrams that can be distinguished from each other by packet types such as packet priority (TOS), data length or protocol type. Traffic 100 is a packet group of “ToS = 5” and traffic 110 is a packet. It consists of a packet group of "TL = 1500 bytes or more". In the example of FIG. 5, the traffic 100 to the traffic 120 output from the terminal 10 is the terminal 10 → router 3
The route is 0 → router 40 → router 50 → terminal 20.

Next, step 2 will be described. (2-1) First, the real network recognition means 205 of the management terminal 200 dynamically collects the network topology information and recognizes the existing network topology, and the topology diagram of the real network and the topology table (the real network topology table Create). Figure 7
-1 indicates a network topology diagram of the network taken as an example here (-1 in FIG. 7).
Is a network obtained by omitting the terminal and the management terminal from the network of FIG. 5). An example of a topology table in this network is shown in -2. Since the cost information is not considered here, "X" indicating an arbitrary value is displayed in the table. -In the example of the topology table in the existing network, the "valid / invalid" parameters of all the inter-router links are set to valid. (2-2) Next, the packet type information setting means 202 of the management terminal 200 sets the packet type for the virtual network. In this example, a packet type 1 for identifying the traffic 100 and a packet type 2 for identifying the traffic 110 are set. An example of the set packet type information is shown in FIG. Referring to FIG. 6, a packet having “TOS = 5” as the packet type information is associated with the virtual network number 1 as the packet type 1, and a packet having “TL = 1500 bytes or more” as the packet type information is virtual as the packet type 2. It is associated with network number 2. (2-3) Next, the inter-router connection information setting means 203 of the management terminal 200 sets valid inter-router connection information (or invalid inter-router connection information) for each packet type. The inter-router connection information indicates whether the connection between the routers in the existing network topology shown in FIG. 7A is valid or invalid as the inter-router connection when performing the route calculation of a specific packet type. Represents In this example, the router-to-router connection information for the packet type 1 is invalid router-to-router connection information that the router-to-router connections between the router 40 and the router 50 and between the routers 40 and 90 are not used for the route calculation of the traffic 100. . The router-to-router connection information for the packet type 2 is invalid router-to-router connection information that the router-to-router connections between the routers 30 and 40 and between the routers 40 and 50 are not used for the route calculation of the traffic 110. (2-4) Next, the virtual network definition means 206 of the management terminal 200, based on the real network topology table created by the real network recognition means 205 and the inter-router connection information created by the inter-router connection information setting means 203, A virtual network topology is constructed for each packet type. In this example, two virtual networks (virtual network 1 and virtual network 2) are defined for packet type 1 and packet type 2. Referring to FIG.
2 shows the topology table of the virtual network 1. In the table, the "valid / invalid" parameter of the inter-router link between the No9 router 40 and the router 50 and the inter-router link between the No10 router 40 and the router 90 is set to "invalid". As a result, in the topology table of the virtual network 1, the two inter-router links are considered not to be connected. -1 shows a topology diagram of the virtual network 1 at this time. Similarly,
2 shows the topology table of the virtual network 2. In the table, "valid / invalid" parameter for the inter-router link between No2 router 30 and router 40 and the inter-router link between No9 router 40 and router 50 is set to "invalid".
Is set to. As a result, in the topology table of the virtual network 2, it is considered that the above two inter-router links are not connected. -1 shows a topology diagram of the virtual network 2 at this time. (2-5) Next, the communication means 204 of the management terminal 200
The packet type information created by the packet type information setting unit 202 and the inter-router connection information for each packet type created by the inter-router connection information setting unit 203 are transmitted to all routers in the network.

Next, step 3 will be described. (3-1) First, the router function unit 101 of the router receives the packet type information and the inter-router connection information from the management terminal 200. (3-2) Next, the packet type information is stored in the packet type table. (3-3) Next, a virtual network topology table for each packet type is created based on the inter-router connection information for each packet type and the actual network topology table, and a virtual routing table for each packet type is created. That is, the first virtual routing table 1031 and the packet type 2 for the packet type 1
To generate a second virtual routing table 1032. At this time, between the router 40 and the router 50 and between the router 40 and the router 90 which are invalid for the packet type 1
The connection between routers between the first virtual routing table 10
The inter-router connection between the router 30 and the router 40 and between the router 40 and the router 50 which is not used for the packet 31 and is invalid for the packet type 2 is not used for the second virtual routing table 1032. (3-4) Next, the association between the generated routing table and the packet type is stored in the packet type table. That is, the packet type table is composed of the IP header information, the virtual network number, and the virtual routing table number.

Next, step 4 will be described. (4-1) The router performs packet relay processing. Referring to FIG. 8, when the router receives the IP packet, the router extracts the header information of the IP packet. Next, it is searched whether the IP header information is registered in the packet type table 104. IP in the packet type table 104
The header information and the condition match (packet type table 10
If there is a virtual network associated with the virtual network, a virtual routing table corresponding to the virtual network is determined and the IP packet is routed using this virtual routing table.
If there is no entry in the packet type table 104 (a packet that does not require route control by the virtual network) or if there is an entry in the packet type table 104 but is not associated with the virtual network, the actual routing table 102 is used. Do the routing. When the corresponding packet is unreachable when the route is calculated in the virtual routing table, the real routing table 102 is used for routing. Regarding the management information packet transmitted from the management terminal 200, the actual routing table 102
Use to route. In this way, the output port for transmitting the received IP packet is determined, and I
Send a P packet. In this example, traffic 100
(Packet group having “TOS = 5” in the IP header information) is recognized as packet type 1 and is routed using the first virtual routing table 1031, and traffic 110 (IP header information has “TL = 1500 bytes or more The packet group having “” is recognized as a packet type 2 and is routed using the second virtual routing table 1032, and the traffic 120 is routed using the real routing table 102. A passage route of the traffic 100 is shown in FIG. The route of the traffic 100 is changed as follows: terminal 10 → router 30 → router 40 → router 70 → router 50 → terminal 20. The dotted line indicates an invalid connection between routers. The passage route of the traffic 110 is shown in FIG.
The route of the traffic 110 is changed as follows: terminal 10 → router 30 → router 40 → router 50 → terminal 20. The dotted line indicates an invalid connection between routers. The passage route of the traffic 120 is the same as in FIG.
0 → router 30 → router 40 → router 50 → terminal 20. The passage routes of the traffic 100 to the traffic 120 are summarized as shown in FIG.

In the above embodiment, the router generates the virtual routing table for each packet type based on the connection information between routers for each packet type and the real network topology table. The router may generate the virtual routing table based on the virtual network topology table by transmitting the network topology table.

In this way, by explicitly defining a plurality of virtual networks according to the packet type on the network, the packets related to the virtual network are routed within the defined route group. Flexible route control suitable for the characteristics can be performed. Further, since the route calculation is dynamically performed within the defined route group, it is possible to follow the topology change and the route change due to the route failure. As a result, there is an effect that the load in the network is distributed and congestion is prevented in advance.

Next, a second embodiment of the present invention will be described in detail with reference to the drawings. FIG. 12 is a diagram showing a configuration of the second exemplary embodiment of the present invention, and represents a network topology (a virtual network topology described later is hereinafter referred to as an actual network topology). FIG. 13 is a diagram showing an example of a virtual network. FIG. 14 is a diagram showing an example of a virtual network.

Referring to FIG. 12, the second embodiment of the present invention has four LAN segments (LAN 300, L
AN310, LAN320 and LAN330) and five routers (router 40) connecting them.
0, router 410, router 420, router 430 and router 440).

Consider a case where communication is performed between LAN segments in the configuration of FIG. As an example, LAN30
When data is transferred from the terminal 0 to the terminal of the LAN 330, considering the passage route so that no loop exists, the following four passage route candidates can be considered. The first is router 400 → router 420 → router 440. The second is router 400 → router 420 → router 430.
→ router 440. As the third, router 400 → router 410 → router 420 → router 440. As the fourth,
Router 400 → Router 410 → Router 420 → Router 4
30 → router 440.

Here, in the management terminal, a virtual network is defined as in the first embodiment. In this example,
Two packet types are set for the network of FIG. 12, connection information between routers is set for each packet type, and two virtual networks shown in FIGS. 13 and 14 are defined.

As a result, from the terminal of LAN 300 to LA
The candidates for the transit route that transfers data to the N330 terminal are
Although four in FIG. 12, FIG. 13 and FIG.
Then, there is only one each.

Packets are distributed to the virtual network of either FIG. 13 or FIG. 14 according to a specific packet type, and the routing calculation is executed in each virtual network.

In each virtual network, the number of route candidates decreases, so that the route calculation process can be reduced.

In this way, by expressing a complicated network configuration by a plurality of simple virtual network configurations, it is possible to reduce the processing of route calculation by the routine protocol, and even for a structurally complex network. It becomes possible to deal with it easily.

For example, in a link state protocol such as OSPF, when there are a plurality of arrival routes of the data flow to be routed, there is a problem that the processing of the route calculation increases, but the route control method according to the present invention is used. By introducing it, this problem can be solved.

[0061]

As a first effect, it is possible to explicitly define a plurality of virtual networks according to packet types on the network, and it is possible to provide a flexible route control method suitable for the characteristics of the network. Is.

The second effect is structurally congestion, propagation delay,
You can build a network that is less likely to drop packets,
In addition, traffic load can be distributed according to packet types that are easy to implement.

The third effect is that, in a specific network, it is possible to define a packet type that matches the characteristics peculiar to the network.

The fourth effect is that it is possible to perform QoS control on a route in which traffic load is distributed by using a traffic engineering method such as Diffserv or RSVP.

A fifth effect is that it is possible to reduce the processing of the route calculation, and it is possible to easily handle the routing protocol even in a structurally complicated network.

The reason is that the packet type information for identifying a packet is set to define a virtual network for each packet type, and when the packet is received, the packet type is identified and the virtual network is routed to the corresponding virtual network. Is.

[Brief description of drawings]

FIG. 1 is a diagram showing a configuration of a first embodiment of the present invention.

FIG. 2 is a diagram showing a configuration of a router

FIG. 3 is a diagram showing a configuration of a management terminal.

FIG. 4 is a diagram showing an example of a packet type table.

FIG. 5 is a diagram explaining the operation of the first embodiment of the present invention.

FIG. 6 is a diagram showing an example of a packet type table.

FIG. 7 is a diagram showing an example of a virtual network topology.

FIG. 8 is a diagram showing a flow of IP packet routing in a router.

FIG. 9 is a diagram showing a modified transit route in traffic 100.

FIG. 10 is a diagram showing modified transit routes for traffic 110.

FIG. 11 is a diagram showing a traffic passage route.

FIG. 12 is a diagram showing a configuration of a second exemplary embodiment of the present invention.

FIG. 13 is a diagram showing an example of a virtual network.

FIG. 14 is a diagram showing an example of a virtual network.

FIG. 15 is a diagram showing an example of an IPv4 header format.

FIG. 16 is a diagram showing an example of an IPv6 header format.

[Explanation of symbols]

10 terminals 20 terminals 30 routers 40 routers 50 routers 60 routers 70 router 80 router 90 router 101 router function part 102 real routing table 104 packet type table 105 Interface part 200 management terminal 201 control means 202 packet type information setting means 203 Router connection information setting means 204 Communication means 205 Real network recognition means 206 virtual network definition means 300 LAN 310 LAN 320 LAN 330 LAN 400 routers 410 router 420 router 430 router 440 router 1031 First virtual routing table 1032 Second virtual routing table

Claims (10)

[Claims] 1. A connectionless packet network (IP) that dynamically collects network topology information using a routing protocol and selects the shortest path.
Traffic load balancing method in a network), comprising: means for defining a plurality of virtual networks for the network; and means for distributing traffic to the virtual network. 2. A connectionless packet network (IP) that dynamically collects network topology information using a routing protocol and selects the shortest route.
Traffic load balancing system in a network), the network includes a terminal for transmitting and receiving packets, a router for relaying packets between the terminals, and a management terminal, the management terminal providing packet type information for identifying the packets. Setting means, means for setting valid / invalid information of each inter-router connection for constructing a virtual network for each packet type, and the packet type information and inter-router connection information for each packet type to the router. Transmitting means, the router generates a real network topology table corresponding to an existing network, and creates a real routing table based on the real network topology table; the packet type information and the packet; Management of connection information between routers for each type Means for receiving from the end, a virtual network topology table for each packet type based on the real network topology table and router-to-router connection information for each packet type, and a virtual routing table based on the virtual network topology table Means for searching and whether the packet received when the packet is received corresponds to the packet type, and if the packet type corresponds to the packet type, the packet is received using the virtual routing table corresponding to the applicable packet type. And a means for routing packets, the traffic load balancing method. 3. A connectionless packet network (IP) that dynamically collects network topology information using a routing protocol and selects the shortest route.
Network), the network comprises a terminal for transmitting and receiving packets, a router for relaying packets between the terminals, and a management terminal, the management terminal collecting network topology information from the network. Network recognition means for recognizing existing network topology and creating a real network topology table, means for setting packet type information for identifying packets, and connection between routers for constructing a virtual network for each packet type Means for setting valid / invalid information of the packet, virtual network defining means for creating a virtual network topology table for each packet type based on the inter-router connection information and the real network topology table, and the packet type information and And means for transmitting a virtual network topology table for each packet type to the router, the router generating an actual network topology table corresponding to an existing network, and performing actual routing based on the actual network topology table. Means for creating a table, means for receiving the packet type information and a virtual network topology table for each packet type from the management terminal, and a virtual network for each packet type based on the virtual network topology table for each packet type Means for creating a virtual routing table corresponding to the above, and when the packet is received, it is searched whether or not the received packet corresponds to the packet type, and if it corresponds to the packet type, the corresponding packet type is supported. Virtual and means for routing the received packet using the routing table, the traffic load balancing method, characterized in that the. 4. The router comprises means for routing the received packet using an actual routing table corresponding to the existing network when the received packet does not correspond to the packet type. The traffic load balancing system according to claim 2 or 3. 5. The traffic load balancing system according to claim 2, wherein the packet type information is composed of a combination of packet header information. 6. A connectionless packet network (IP) that dynamically collects network topology information using a routing protocol and selects the shortest route.
Traffic load balancing method in a network), in which the router collects network topology information from the network, generates a real network topology table corresponding to an existing network, and creates a real routing table based on the real network topology table. The management terminal sets packet type information for identifying a packet, sets inter-router connection information for constructing a virtual network for each packet type, and connects the packet type information and the router for each packet type. The connection information is transmitted to each router that constitutes the network, and the router receives the packet type information and the inter-router connection information for each packet type from the management terminal, and the real network topology table and the packet type. A virtual network topology table for each packet type is created based on the inter-router connection information, and a corresponding virtual routing table is created based on the virtual network topology table. The router receives the packet when it receives the packet. A traffic characterized by searching whether or not a packet corresponds to the packet type, and if the packet type corresponds to the packet type, the received packet is routed using a virtual routing table corresponding to the applicable packet type. Load balancing method. 7. A connectionless packet network (IP) which dynamically collects network topology information by using a routing protocol and selects a shortest route.
Traffic load balancing method in a network), in which the router collects network topology information from the network, generates a real network topology table corresponding to an existing network, and creates a real routing table based on the real network topology table. The management terminal collects network topology information from the network, recognizes the existing network topology, creates a real network topology table, sets packet type information for identifying packets, and sets a virtual network for each packet type. Setting the inter-router connection information for building the, to create a virtual network topology table for each packet type based on the inter-router connection information and the actual network topology table, The packet type information and the virtual network topology table for each packet type are transmitted to each router forming the network, and the router receives the packet type information and the virtual network topology table for each packet type from the management terminal. Creating a virtual routing table corresponding to the virtual network for each packet type based on the virtual network topology table for each packet type,
When the router receives a packet, the router searches the packet to see if it matches the packet type, and if it does, receives the packet using the virtual routing table corresponding to the packet type. A traffic load balancing method characterized by performing packet routing. 8. The received packet is routed using an actual routing table corresponding to an existing network when the received packet does not correspond to the packet type. Traffic load balancing method. 9. The packet type information is a combination of packet header information.
Alternatively, the traffic load balancing method according to item 7. 10. A real routing table corresponding to an existing network when the received packet corresponds to the packet type and the destination is unreachable using the virtual routing table of the corresponding virtual network. The traffic load balancing method according to claim 6 or 7, wherein the received packet is routed using the.