JP3836853B2 - Load balancing communication network, subscriber user accommodation node, and network control server - Google Patents

Description

  The present invention relates to a route control technique for a packet communication network, and more particularly to traffic engineering that realizes high reliability and optimization of a network by controlling the flow of traffic in the network.

Conventionally, in IP networks represented by the Internet, as a data communication service between subscriber users, a best-effort communication service that routes the subscriber users through the shortest route and does not guarantee throughput when the network is congested is provided. Has been.
Best-effort communication services ensure the reachability while suppressing the consumption of network resources by selecting the minimum necessary shortest route as the route connecting subscriber users with the shortest route-based routing. .

However, in such a shortest route-based routing, traffic is not taken into consideration when selecting a route. For example, there is a possibility that a large number of user IP flows may be concentrated on an arbitrary route and congestion may occur. Communication reliability and communication quality were low.
As a technique for solving this problem, traffic engineering for realizing high reliability and optimization of the network by controlling the flow of traffic in the network has been studied.

In an IP network employing such traffic engineering, a transfer route is assigned to data communication between subscriber users according to the band used. That is, in each frame forwarding node on the IP network, a plurality of forwarding routes are prepared in advance for the same destination, and when congestion and a failure are detected in a specific forwarding route, the route is replaced with the alternative route. A frame storing user data is transferred.
With this technology, traffic can be distributed to each path, and the reliability and communication quality for data communication can be improved.

The applicant has not yet found prior art documents related to the present invention by the time of filing other than the prior art documents specified by the prior art document information described in this specification.
Yuuichi Naruse et, al. "IP-in-IPv6 0verlay Networking Technology for a Terabit Class Super-network", NTT Technical Review, VOL.2, NO.3 Mar. 2004 2003 Kenichi Matsui et, al. "Cut through Optical Path Control Technology for a Terabit Class Super-network", NTT Technical Review, VOL.2, NO.3 Mar. 2004 2003

However, in such conventional traffic engineering, when assigning a transfer path for data communication between subscriber users, only the bandwidth used for each data communication is taken into account. There was a problem that it could not be obtained.
For example, in a situation where a plurality of users simultaneously use a plurality of applications, an IP flow with a relatively short frame length such as VoIP (Voice over IP) and an IP flow with a relatively long frame length are mixed. When the traffic amount is uniformly allocated to a plurality of routes by the conventional traffic engineering, there is a bias that user frames having a short frame length concentrate on any of the routes.

  At this time, since the processing load on the processor of the frame transfer apparatus is affected not only by the traffic amount but also by the increase / decrease of the number of frames, the frame transfer apparatus installed on the path has a larger number than the other paths. Frames may be processed, and the number of frames exceeding the processing capacity of the processor may be received. Therefore, when the processing load becomes larger than the processing capacity of the processor, the frame is discarded, and as a result, there is a problem that communication between users is affected.

  The present invention is to solve such problems, and provides a load distribution type communication network, a subscriber user accommodation node, and a network control server that can obtain high reliability for data communication between subscriber users. The purpose is to do.

  In order to achieve such an object, the communication network according to the present invention uses a plurality of transfer paths each having a unique path identifier between subscriber user accommodation nodes accommodating a plurality of subscriber users. This is a communication network in which a backbone network is configured by connecting to a network, and when a frame storing user data is transferred between subscriber users, the transfer route of the frame is clearly indicated using a path identifier at the subscriber user accommodation node. In the subscriber user accommodation node, a forwarding table for deriving a destination subscriber user accommodation node and a route identifier of the output destination forwarding route for the destination user address, and a subscriber user accommodation in which a frame received from the subscriber user is different from the subscriber user accommodation node. When sending to the node, the transfer table is sent according to the destination user address of the frame. Based on the path identifier obtained from the output destination transfer path in which and a transfer control unit for determining a.

  Then, the total number of transfer frames transferred per unit time of each frame having the same path identifier as the destination user address, the destination subscriber user accommodation node, and the number of transfer frames for each pair of the subscriber user accommodation node and the path identifier Counter table for managing the total value obtained in this manner, and transfer of the corresponding entry in the counter table based on the destination user address, destination subscriber user accommodation node, and route identifier extracted from the frame at the time of frame transfer A frame count unit that updates the number of frames and the total value respectively, and when the total value of the number of transfer frames for any pair of subscriber user accommodation nodes and path identifiers exceeds a threshold in the counter table, the path identifier is excluded. One of the route identifiers addressed to the subscriber user accommodation node is A feedback control unit that selects from the list, and the transfer control unit performs feedback control on an arbitrary entry of a destination user address to which a route identifier whose total value exceeds a threshold is assigned in the transfer table. The route identifier selected in the section is assigned.

  At this time, the number of frames transferred per unit time is collected for each transfer path from the frame transfer node that is provided in the backbone network and transfers frames between the transfer paths. If there is a difference, a network control server for changing and controlling the forwarding table in the subscriber user accommodation node may be further provided so that the difference is reduced.

Further, instead of the number of transfer frames, it may be used load value obtained by synthesizing the total frame length transferred per count and unit time transfer frame transferred per unit of time.

  Further, when selecting a route identifier by the feedback control unit, a pair of route identifiers having a minimum total number of transfer frames among the pairs of subscriber user accommodation nodes and route identifiers managed in the counter table is selected. You may make it select.

  Further, a network control server may be further provided that monitors a failure in the backbone network and, when a failure is detected, identifies a transfer path affected by the failure and notifies the subscriber user accommodation node.

  At this time, the transfer control unit may select a transfer route different from the transfer route affected by the failure notified from the network control server as the transfer route used for transferring the frame at the time of frame transfer. .

  Further, a plurality of virtual private networks that connect each subscriber user accommodation node may be provided in the backbone network, and a physically different transfer path may be set for each virtual private network.

  In addition, based on the physical structure of the backbone network and the number of virtual private networks, a calculation is performed to allocate a physically different transfer route for each virtual private network, and the transfer is provided in the backbone network according to the calculation result. A network control server for setting a physically different transfer route for each virtual private network by remotely setting a transfer table of the frame transfer node for a frame transfer node that transfers frames between routes may be further provided. Good.

  In addition, when the transfer control unit transmits the user network format frame received from the accommodated subscriber user to the backbone network, the frame may be encapsulated into a predetermined backbone network format frame.

  The subscriber user accommodation node according to the present invention connects the subscriber user accommodation nodes that accommodate a plurality of subscriber users in a network using a plurality of transfer paths each having a unique path identifier. Therefore, when a frame in which a user data is stored is transferred between subscriber users by configuring a backbone network, a subscriber user used in a communication network that clearly indicates a frame transfer path using a path identifier at a subscriber user accommodation node A receiving node that transmits a destination subscriber user accommodating node corresponding to a destination user address and a route identifier of an output destination forwarding route and a frame received from the subscriber user to a subscriber user accommodating node different from itself. Is obtained from the forwarding table according to the destination user address of the frame. In which and a transfer control unit that determines an output destination transfer path based on the path identifier.

  Then, the total number of transfer frames transferred per unit time of each frame having the same path identifier as the destination user address, the destination subscriber user accommodation node, and the number of transfer frames for each pair of the subscriber user accommodation node and the path identifier Counter table for managing the total value obtained in this manner, and transfer of the corresponding entry in the counter table based on the destination user address, destination subscriber user accommodation node, and route identifier extracted from the frame at the time of frame transfer A frame count unit that updates the number of frames and the total value respectively, and when the total value of the number of transfer frames for any pair of subscriber user accommodation nodes and path identifiers exceeds a threshold in the counter table, the path identifier is excluded. One of the route identifiers addressed to the subscriber user accommodation node is A feedback control unit that selects from the list, and the transfer control unit performs feedback control on an arbitrary entry of a destination user address to which a route identifier whose total value exceeds a threshold is assigned in the transfer table. The route identifier selected in the section may be assigned.

In this case, instead of the number of transfer frames, it may be used load value obtained by synthesizing the total frame length transferred per count and unit time transfer frame transferred per unit of time.

  Further, when selecting a route identifier by the feedback control unit, a pair of route identifiers having a minimum total number of transfer frames among the pairs of subscriber user accommodation nodes and route identifiers managed in the counter table is selected. You may make it select.

  In addition, when the transfer control unit transmits the user network format frame received from the accommodated subscriber user to the backbone network, the frame may be encapsulated into a predetermined backbone network format frame.

  In addition, the network control server according to the present invention connects the subscriber user accommodation nodes accommodating a plurality of subscriber users in a network using a plurality of transfer paths each having a unique path identifier. When a frame containing user data is transferred between subscriber users in a backbone network, the subscriber user node is used in a communication network that clearly indicates a frame transfer path using a path identifier at a subscriber user accommodation node. The number of frames transferred per unit time for each transfer path from a frame transfer node that is provided in the backbone network and transfers frames between transfer paths. A traffic frame between the traffic information transfer unit and the transfer path If there is a difference between the is obtained by providing a load distribution calculation unit that changes control the forwarding table of the subscriber user accommodation within the node so that the difference becomes smaller.

In this case, instead of the number of transfer frames, it may be used load value obtained by synthesizing the total frame length transferred per count and unit time transfer frame transferred per unit of time.

  In addition, the load distribution calculation unit may monitor a failure in the backbone network, and when a failure is detected, a transfer path affected by the failure may be specified and notified to the subscriber user accommodation node.

  In addition, the load balancing calculation unit performs a calculation to allocate a physically different transfer route for each virtual private network based on the physical structure of the backbone network and the number of virtual private networks, and the backbone network is determined according to the calculation result. For a frame forwarding node that is provided inside and forwards a frame between forwarding routes, a forwarding table possessed by the frame forwarding node is remotely set to set a physically different forwarding route for each virtual private network. May be.

  According to the present invention, when the number of transfer frames exceeds a threshold value in an arbitrary transfer route, the transfer table is controlled, and the transfer is performed for an arbitrary entry of a destination user address using the transfer route. Since a route identifier other than the route identifier of the route is assigned, traffic engineering considering the number of transfer frames transferred per unit time can be realized.

  Thus, for example, in a situation where an IP flow with a relatively short frame length and an IP flow with a relatively long frame length are mixed, such as VoIP due to multiple users simultaneously using multiple applications, When traffic is uniformly allocated to multiple paths based on the bandwidth used as in engineering, there is a possibility that frames with short frame lengths will be concentrated on any of the paths. For example, it is possible to accurately grasp the processing load on the processor of the frame forwarding node located on each path, and it is possible to reliably avoid such a bias.

  Therefore, the processing load on the processor of each frame forwarding node can be equalized more accurately, and further, the route switching at the time of congestion or failure can be performed quickly. It is possible to realize a load distribution type communication network that can obtain high reliability for data communication.

Next, embodiments of the present invention will be described with reference to the drawings.
[First Embodiment]
First, a load distribution type communication network according to a first embodiment of the present invention will be described with reference to FIG. FIG. 1 is a block diagram showing a network model of a load distribution type communication network according to a first embodiment of the present invention.
This communication network includes subscriber user accommodation nodes 1 to 4, frame transfer apparatuses 6 to 9, and a network control server 5 that controls them.

The terminals 10 to 17 are accommodated in the subscriber user accommodation nodes 1 to 4 via the user networks 18 to 21. The subscriber user accommodation nodes 1 to 4 are connected by frame transfer devices 6 to 9.
A backbone network composed of these subscriber user accommodation nodes 1 to 4 and frame transfer devices 6 to 9 is referred to as a core network 22, and a network including terminals 10 to 17 is referred to as a user network 23.

[Physical model]
Next, with reference to FIG. 2, a physical model of the load balancing communication network according to the first embodiment of the present invention will be described. FIG. 2 is an example of a physical model of the load distribution type communication network according to the first embodiment of the present invention.

As shown in the physical model example of FIG. 2, the subscriber user accommodation nodes 1 to 4 are connected using a plurality of transfer paths.
For example, the subscriber user accommodation node 1 and the subscriber user accommodation node 3 are configured such that a transfer route including a link 101-frame transfer device 6-link 109, a transfer route including a link 102-frame transfer device 7-link 110, a link A transfer path consisting of 103-frame transfer apparatus 8-link 111 and a transfer path consisting of link 104-frame transfer apparatus 9-link 112 are connected in parallel.

At this time, a route identifier for identifying the route is assigned in advance to each route.
For example, in this embodiment, a path identifier R1 is assigned to a transfer path composed of link 101-frame transfer apparatus 6-link 109. Similarly, a path identifier R2 is assigned to the transfer path consisting of link 102-frame transfer apparatus 7-link 110, and a path identifier R3 is assigned to the transfer path consisting of link 103-frame transfer apparatus 8-link 111. A path identifier R4 is assigned to a transfer path including the link 104, the frame transfer apparatus 9, and the link 112.

The network control server 5 is connected to each node and each device in the core network 22 via links 117 to 124.
That is, the network control server 5 is connected to the subscriber user accommodation nodes 1 to 4 via the links 117 to 120, and is connected to the frame transfer apparatuses 6 to 9 via the links 121 to 124, respectively.

  In the present embodiment, these links 117 to 124 ensure connectivity with each device. This makes it possible to avoid congestion and failure due to physical causes such as link breaks and device congestion by switching the output destination transfer path. Further, centralized management type and centralized control type traffic engineering by the network control server 5 can be realized. Furthermore, centralized management of the network state by the network control server 5 can be realized, and as a result, it is possible to detect a failure and specify all paths affected by the failure.

[Logical model]
Next, a logical model of the load distribution type communication network according to the first embodiment of the present invention will be described with reference to FIG. FIG. 3 is a logical model example of the load distribution type communication network according to the first embodiment of the present invention.

  In the load distribution type communication network according to the first embodiment of the present invention, a plurality of virtual private networks that connect all the subscriber user accommodation nodes are constructed in the backbone network. It is conceivable to set a different transfer route for each.

In the logical model example of FIG. 3, the transfer path via the frame transfer apparatus 6 is configured by the virtual private network VPN # 1, the transfer path via the frame transfer apparatus 7 is configured by the virtual private network VPN # 2, and the frame transfer apparatus 8 The transfer path via the virtual private network VPN # 3 is configured, and the transfer path via the frame transfer device 9 is configured with the virtual private network VPN # 4.
As a result, even when congestion or failure due to a physical cause such as link breakage or device congestion occurs, it is possible to avoid the failure by switching the output destination virtual private network.

[Subscriber user accommodation node]
Next, with reference to FIG. 4, the subscriber user accommodation node used in the load distribution type communication network according to the first embodiment of the present invention will be described in detail. FIG. 4 is a block diagram showing a configuration of a subscriber user accommodation node used in the load distribution type communication network according to the first embodiment of the present invention.

  The subscriber user accommodation nodes 1 to 4 include an input interface unit (hereinafter referred to as an input IF unit) 25, a transfer control unit 26, a load observation unit 27, an output interface unit (hereinafter referred to as an output IF unit) 28, and a server interface unit. (Hereinafter referred to as a server IF unit) 29.

[Input IF section]
The input IF unit 25 has a viability response unit 30. This survivability response unit 30 has a function of returning a Ping response to the transmission source when receiving an Internet Control Message Protocol (ICMP) Ping response request packet from another subscriber user accommodation node. .

  When the header area of the received frame is a core address, the input IF unit 25 determines that the frame is a core frame exchanged on the core network 22, decapsulates and extracts a user frame, and extracts the user frame When the function of transferring a frame to the transfer control unit 26 and the header area of the received frame is a user address, it is determined that the frame is a user frame exchanged on the user network and transferred to the transfer control unit 26 And a function of returning a Ping response to a Ping response request from another subscriber accommodation node using the survivability response unit 30.

[Transfer control section]
The transfer control unit 26 has a transfer table 31. The transfer table 31 includes a destination specifying table 32 and a route specifying table 33.
The destination specifying table 32 has a function of specifying a subscriber user accommodating node accommodating the destination user from the destination user address of the frame.
The route specification table 33 has a function of deriving an output destination route identifier from the destination user address of the frame.

In the core network 22, transfer is performed using a core address composed of an address of a destination subscriber user accommodation node and a path identifier.
The transfer control unit 26 specifies the address of the subscriber user accommodating node accommodating the destination user from the destination user address of the received user frame. On the other hand, when the address of the specified subscriber user accommodation node is not the address that the user has, the route identifier of the output destination is determined from the destination user address of the received user frame, and the address of the identified subscriber user accommodation node is determined. Then, a destination core address is determined from the path identifier, a core header is generated using the address of the core address as a transmission source address, and the user header is encapsulated in the user frame by adding the core header to the user frame. Send to.

On the other hand, when the specified subscriber user accommodation node is itself, the transfer control unit 26 transmits the user frame to the output IF unit 28.
In addition, when the transfer control unit 26 is instructed by the feedback control unit 37 of the load observation unit 27 described later to change the route identifier for an arbitrary destination user address described in the route specification table 33, the transfer control unit 26 The route specifying table 33 is searched using the destination user address thus set as a search key, and the route identifier of the hit entry is rewritten to the route identifier specified by the feedback control unit 37. This rewriting can be realized by adding and deleting entries.

In addition, the transfer control unit 26 has a function of returning an SNMP reference response when receiving an SNMP (Simple Network Management Protocol) reference request of the destination specifying table 32 and the route specifying table 33 from the network control server 5. Yes.
When the transfer control unit 26 receives an SNMP setting request for changing the address of the destination subscriber user accommodation node described in the destination specification table 32 from the network control server 5, the transfer control unit 26 sets the corresponding entry according to the request. Add and delete the address of the destination subscriber user accommodation node. Similarly, when an SNMP setting request for changing the route identifier described in the route specification table 33 is received from the network control server 5, the route identifier of the corresponding entry is added and deleted according to the request.

[Load observation section]
The load observation unit 27 includes a frame count unit 34, a total frame length observation unit 35, a load value calculation unit 36, and a feedback control unit 37.
The frame count unit 34 has a counter table 38.
The counter table 38 has a function for deriving the number of passing frames for the destination user address and the destination core address, and a function for deriving the total number of passing frames for the destination core address.

  The frame count unit 34 counts the number of passing frames for the destination user address and the destination core address, and simultaneously counts the total number of passing frames for the destination core address. When the total number of passing frames for any destination core address exceeds the threshold, the address of the destination subscriber user accommodation node described in the destination core address whose count value exceeds the threshold is the same, and the path identifier Only the destination core addresses that are different from each other, the destination user addresses described in the counter table 38 together with them, and the count value for the pair are transmitted to the feedback control unit. Note that threshold values are individually provided for each entry.

The total frame length observation unit 35 has a band observation table 39.
This band observation table 39 has a function for deriving the total value of the frame lengths of the passing frames for the destination user address and the destination core address, and a function for deriving the total value of the total frame length for the destination core address.

  The total frame length observation unit 35 counts the total value of the frame lengths of the passing frames for the destination user address and the destination core address, and simultaneously counts the total value of the total frame lengths for the destination core address. When the total value of the total frame lengths for any destination core address exceeds the threshold value, the address of the destination subscriber user accommodation node described in the destination core address whose count value exceeds the threshold value is the same. The destination core addresses having different path identifiers only, the destination user addresses described in the band observation table 39 together with these destination core addresses, and the total value of the frame lengths for these pairs are transmitted to the feedback control unit. Note that threshold values are individually provided for each entry.

The load value calculation unit 36 has a load value management table 40.
The load value management table 40 has a function of deriving a load value for the destination user address and the destination core address, and a function of deriving a total load value for the destination core address.

  The load value calculation unit 36 periodically refers to the frame count unit 34 and the total frame length observation unit 35 to obtain the total number of passing frames and the total frame length of the passing frames for the destination user address and the destination core address. And calculating the load value for the destination user address and the destination core address and describing them in the load value management table 40, and the load for each destination core address. A function of summing the values and describing them in the load value management table 40.

  Also, when the total load value for any destination core address exceeds the threshold value, the address of the destination subscriber user accommodation node described in the destination core address whose load value exceeds the threshold value is the same and only the route identifier Each of the different destination core addresses, the destination user addresses described in the load value management table 40, and the load value for the pair are transmitted to the feedback control unit. Note that threshold values are individually provided for each entry.

  The feedback control unit 37 is described from the frame count unit 34, the total frame length observation unit 35, and the load value calculation unit 36 in the count value, the total value of the frame length, and the destination core address where the load value exceeds the threshold value. Destination core addresses with the same destination subscriber user accommodating node address but different route identifiers, destination user addresses described in each table together with these destination core addresses, and count values for these pairs, total frame length When the value and load value are notified, the destination core address and destination user address pair with the smallest count value, total frame length, and load value is extracted, and the destination subscription described in the destination core address is extracted. The address and route identifier of the user user accommodation node are extracted.

  Then, the route identifier assigned to the pair of the destination user address and the destination subscriber user accommodating node whose count value, total frame length value, and load value exceed the threshold is assigned to the transfer control unit 26 with the same destination user. It is instructed to change to a path identifier having a pair of address and destination subscriber user accommodation node and having the smallest count value, total frame length, and load value. At the same time, in the counter table 38, the bandwidth observation table 39, and the load value management table 40, a pair of a destination user address and a destination subscriber user accommodation node whose count value, total frame length, and load value exceeds a threshold value. The route identifier assigned to is reset to the changed one.

The load observation unit 27 validates any one of data of the frame count unit 34, data of the total frame length observation unit 35, and data of the load value calculation unit 36 as data to be input to the feedback control unit 37. And has the function of preventing other data from being input to the feedback control unit 37.
In addition, when receiving an SNMP reference request for the counter table 38, the bandwidth observation table 39, and the load value management table 40 from the network control server 5, it has a function of returning an SNMP reference response and simultaneously resetting the counter value of each table. is doing.

[Output IF section]
The output IF unit 28 includes an output link identification table 41 and a viability confirmation unit 42. The output link specification table 41 has a function of specifying the output link from the destination address described in the header area of the frame.

If the header area is a core header, the frame is output to the core network. If the header area is a user header, the frame is output to the user network.
The viability confirmation unit 42 includes a viability confirmation table 43. The survivability confirmation table 43 has a function of deriving a timer and a flag indicating survivability confirmation from the address and path identifier of each subscriber user accommodation node.

The survivability confirmation unit 42 uses each transfer route to each subscriber user accommodation node to periodically transmit a Ping response request and manage whether or not the Ping response has been received regularly. And a function of notifying the network control server 5 of the corresponding subscriber user accommodation node and route identifier as a failure notification when a periodic Ping response cannot be confirmed.
The output IF unit 28 has a function of outputting a packet to each link and a function of periodically confirming route reachability of each transfer route.

[Server IF section]
When the server IF unit 29 receives the SNMP reference request and the SNMP setting request from the network control server 5, the server IF unit 29 identifies whether the transmission destination of the request is the transfer control unit 26 or the load monitoring unit 27. And a function for transferring the response to the network control server 5 when an SNMP request response is received from the transfer control unit 26 and the load observation unit 27.

[Operation of First Embodiment]
Next, the operation of the load distribution type communication network according to the first embodiment of the present invention will be described. Below, the data communication between the subscriber user accommodation node 1 and the subscriber user accommodation node 3 is demonstrated as an example.

In the communication network in this operation example, as shown in FIG. 5, it is assumed that a unique address is assigned to each terminal and subscriber user accommodation node. FIG. 5 is an example of a physical model of a communication network used in the operation example.
The addresses of user # 1 to user # 8 are assigned to the terminals 10 to 17, respectively, and the core # 1 to core # 4 are assigned to the subscriber user accommodation node 1 to the subscriber user accommodation node 4, respectively. It has been.
The subscriber user accommodation node 3 accommodates a terminal 50 via the user network 20, and an address of user # 9 is assigned to the terminal 50.

Further, as shown in FIG. 6, it is assumed that a unique route identifier is assigned to each route connecting the subscriber user accommodation nodes 1 to 4. FIG. 6 is a logical model example of the communication network used in the operation example.
The route identifier R1 is assigned to the route via the frame transfer device 6, and the route identifier R2 is assigned to the route via the frame transfer device 7. A route identifier R3 is assigned to the route via the frame transfer device 8, and a route identifier R4 is assigned to the route via the frame transfer device 9.

[Operation of subscriber user accommodation node]
The operation of the subscriber user accommodation node 1 will be described with reference to FIG. FIG. 7 is an explanatory diagram showing the operation of the subscriber user accommodation node 1. Hereinafter, a case where communication is started between the terminal 10 and the terminal 14 will be described as an example.
Communication between the terminal 10 and the terminal 14 is user flow A, and a user frame 201 shown in FIG. 8 is used. FIG. 8 is a configuration example of a user frame and a core frame.
As shown in FIG. 8, the user frame 201 includes a user head and a payload. When used in the user flow A, the user head 201 includes a user # 1 indicating a transmission source user address and a destination user address. Is stored.

The subscriber user accommodation node 1 receives the user frame 201 from the terminal 10 at the input IF unit 25, determines that the frame is the user frame 201, and transfers the frame to the transfer control unit 26.
The transfer control unit 26 refers to the destination specifying table 32 shown in FIG. 9 and specifies the core # 3 that identifies the subscriber user accommodating node 3 that is the transfer destination from the user # 5 that is the destination user address of the user frame 201. To do. FIG. 9 is a configuration example of the destination specifying table 32, and each destination subscriber user accommodating node is associated with each destination user address.

Thereafter, the route identification table 33 shown in FIG. 10 is referred to, and the route identifier R1 of the route used for the transfer of the user frame 201 is derived from the user # 5 that is the destination user address of the user frame 201. FIG. 10 is a configuration example of the route specification table 33, and each destination user address is managed in association with each route identifier.
As a result, the destination core address core # 3_R1 is calculated for the user frame 201, and a core header composed of the destination core address core # 3_R1 and the source core address # 1 is given to the user frame 201. The core frame 202 shown in FIG. 8 is generated.

As shown in FIG. 8, the core frame 202 includes a core header and a user frame 201. In this case, the core header includes a core # 1 indicating a source core address and a core # 3_R1 indicating a destination core address. Is stored.
The core frame 202 generated in this way is transferred to the load observation unit 27.
When the own core address core # 1 is calculated in the path specification table 33, the transfer control unit 26 transfers the corresponding user frame 201 to the output IF unit 28.

  When receiving the core frame 202, the load observation unit 27 extracts the destination user address user # 5 and the destination core address core # 3_R1 from the core frame 202, and the counter value of the corresponding entry in the counter table 38 illustrated in FIG. Is increased by “1” and is increased from “100” to “101”. FIG. 11 is a configuration example of the counter table 38, and the number of transferred frames and the threshold value are associated and managed for each pair of the destination user address and the destination core address.

  Furthermore, when the load observation unit 27 measures the frame length of the core frame 202 and measures 1000 bytes, the load observation unit 27 increases the total frame length of the corresponding entry in the band observation table 39 illustrated in FIG. 12 by 1000 bytes and increases from “212” to “213”. To "". FIG. 12 shows an example of the configuration of the bandwidth observation table 39. For each pair of the destination user address and the destination core address, the total value of the total frame lengths of the transferred frames and their threshold values are managed in association with each other. Yes.

  Then, the load observation unit 27 counts the counter value of the corresponding entry managed in the counter table 38 of the frame count unit 34 and the total frame length of the corresponding entry managed in the band observation table 39 of the total frame length observation unit 35. Are calculated by, for example, calculating the total value of both, and the load value for the pair of the destination user address and the destination core address is calculated and described in the load value management table 40 shown in FIG. FIG. 13 shows a configuration example of the load value management table 40, in which a load value and a threshold value thereof are associated and managed for each pair of a destination user address and a destination core address.

In this operation example, the total value of the counter value and the total frame length is used as the load value. As shown in FIG. 13, the load value for the pair of the destination user address user # 5 and the destination core address core # 3_R1 was “312” before the transfer process, but the counter value “100” and the total frame length “ Since “212” is added by “1”, the load value after the transfer process is “314”.
The load observation unit 27 transfers the core frame 202 to the output IF unit 28 after completing the series of controls described above.

  The output IF unit 28 receives the core frame 202 from the load observation unit 27, refers to the output link specification table 41 shown in FIG. 14, specifies the output link corresponding to the destination address, and sends the core frame to the specified output link. 202 is output. FIG. 14 is a configuration example of the output link identification table 41, and the output link is managed in association with each destination address. In the example of FIG. 14, the output link 101 is specified as the output link corresponding to the core # 3_R1 that is the destination address of the core frame 202, and the core frame 202 is output to the output link 101.

Here, the load observation unit 27 checks the counter table 38 of the frame count unit 34 and validates the data. In the counter table 38 of FIG. 11 described above, the total number of transfer frames of the core # 3_R1 is “5000”. When the corresponding frame is transferred thereafter, the threshold value “5000” is exceeded.
In such a case, the load observation unit 27 extracts the entry having the core # 3 from the entries in the counter table 38. In this example, the core # 3_R2 and its count value “1203”, the core # 3_R3 and its count value “3254”, and the core # 3_R4 and its count value “2782” are extracted. Then, the feedback control unit 37 is notified of these entries and the counter value.

The feedback control unit 37 selects an arbitrary entry, that is, a new destination user address indicating a different transfer path based on each count value from among the entries notified from the load observation unit 27, and is assigned to the core # 3_R1. The transfer control unit 26 is instructed to assign the existing destination user address to the selected new destination user address.
As a method of selecting a new destination user address at this time, there are a method of selecting arbitrarily and a method of selecting a destination user address having an entry with the smallest counter value in the counter table 38. In the latter case, the core # 3_R2 having the smallest counter value “1203” is selected and notified to the transfer control unit 26.

  In response to this, the transfer control unit 26 searches the destination specifying table 32 and the route specifying table 33 using the destination user address instructed from the feedback control unit 37 as a search key, and determines the path identifier of the hit entry as the feedback control unit. The route identifier specified from 37 is rewritten. In the example of FIGS. 9 and 10, the user # 5 is searched for as the destination user address, and the transfer route corresponding to the user # 5 is rewritten from R1 to R2 in the route specifying table 33.

  Thereby, when a frame having user # 5 as the destination user address arrives at the subscriber user accommodation node 1 from each terminal thereafter, the transfer control unit 26 refers to the destination specification table 32 and the route specification table 33. Core # 3_R2 is generated as a destination core address that identifies the transfer route R2 as a route to the destination subscriber user accommodation node 3. Accordingly, for the frame addressed to the user # 5 arriving at the subscriber user accommodation node 1 thereafter, the destination subscriber is avoided via the transfer route R1 having a relatively high load and passing through the transfer route R2 having a relatively low load. It is transferred to the user accommodation node 3.

  The counter table 38, the bandwidth observation table 39, and the load value management table 40 of each subscriber user accommodation node 1 to 4 are periodically referred to from the network control server 5, and these tables are referred to. At this time, the count value, total packet length, and load value of each table are all reset to 0 (zero). The values in these tables may be initialized autonomously by the subscriber user accommodation nodes 1 to 4.

  Thus, in the communication network according to the present embodiment, the number of transfer frames output to each transfer path per unit time is observed, and when the number of transfer frames exceeds a threshold value in an arbitrary transfer path, Since the forwarding table is controlled and a route identifier other than the route identifier of the forwarding route is assigned to any entry of the destination user address that uses the forwarding route, the number of forwarding frames transferred per unit time Traffic engineering that takes into account

  Thus, for example, in a situation where an IP flow with a relatively short frame length and an IP flow with a relatively long frame length are mixed, such as VoIP due to multiple users simultaneously using multiple applications, If traffic volume is uniformly allocated to multiple paths based on the bandwidth used as in engineering, there may be a bias that frames with short frame lengths such as VoIP concentrate on any path. According to this form, it is possible to accurately grasp the processing load on the processor in the frame forwarding node located on each path, and it is possible to reliably avoid such a bias.

  Therefore, the processing load on the processor in each frame forwarding node can be equalized more accurately, and as a result, a load distribution type communication network that can obtain high reliability for data communication between subscriber users is realized. it can.

At this time, instead of the number of transfer frames, the total frame length transferred per unit time of each frame having the same path identifier as the destination user address, the destination subscriber user accommodation node may be used. Traffic engineering that takes into account the total frame length transferred to
Accordingly, it is possible to equalize the link usage efficiency in each path more accurately, and as a result, it is possible to realize a load distribution type communication network that can obtain high reliability for data communication between subscriber users.

Further, instead of the number of transfer frames, a load value obtained by combining the number of frames and the total frame length may be used, and traffic engineering considering both parameters can be realized.
Therefore, the processing load on the processor of the frame forwarding node arranged on each route and the link usage efficiency on each route can be equalized. As a result, the data communication between each subscriber user can be prevented. A load-distributed communication network with high reliability can be realized.

  In addition, when a frame received from a subscriber user is transmitted to a backbone network in a subscriber user accommodation node, a user network format frame is encapsulated into a backbone network format frame. By using together, it becomes possible to collect the communication to be controlled from the user unit to the subscriber user accommodation node unit.

[Second Embodiment]
Next, a load distribution type communication network according to the second embodiment of the present invention will be described with reference to FIG. FIG. 15 is a block diagram showing a configuration of a network control server used in the load distribution type communication network according to the second embodiment of the present invention. The network model, physical model, logical model, and subscriber user accommodating node of the load distribution type communication network according to the second embodiment are described in the first embodiment (FIGS. 1 to 4). The detailed description is omitted here.

As shown in FIG. 15, the network control server 5 includes a traffic information exchange unit 44 and a load distribution calculation unit 45.
The traffic information exchange unit 44 has a traffic information exchange table 46. The traffic information exchange table 46 has a function of deriving an output link and a timer from the addresses of the devices in the network. The initial value of the timer is set by the traffic information transmission / reception unit 44, is counted down according to the time, and is reset to the initial value again when it reaches 0 (zero).

  The traffic information exchanging unit 44 implements SNMP, and the destination identification table 32 and the path identification table 33 of the subscriber user accommodation node of the entry whose timer of the traffic information exchange table 46 becomes 0 (zero) by the SNMP reference request. When receiving an SNMP reference response in which the function of referring to the counter table 38, the bandwidth observation table 39, the load value management table 40, and the output link specifying table 41 and the referenced traffic information is described, the traffic information is extracted. And a function of notifying the load distribution calculating unit 45.

  Furthermore, when a failure notification is received from the survivability confirmation table 43 of the output IF unit 28, a function for notifying the load distribution calculation unit 45 describing failure information later, and a route notified from the load distribution calculation unit 45 described later Based on the information, when adding and deleting entries in the destination identification table 32, the path identification table 33, and the output link identification table 41 of the subscriber user accommodation node, an entry notified from the load distribution calculation unit 45 is set as an SNMP setting request. As a function of transmitting to the subscriber user accommodation node.

The load distribution calculation unit 45 includes a traffic information holding unit 47, a route optimization processing unit 48, and an update route information holding unit 49.
The traffic information holding unit 47 has a function of holding traffic information and failure path information notified from the traffic information exchanging unit 44.
The route optimization processing unit 48 uses the traffic information and the failure route information stored in the traffic information holding unit 47 to execute the load distribution calculation in the network, the entry calculated by the load distribution calculation, and the entry. It has a function of notifying the update path information holding unit 49 of the subscriber user accommodation node to be set and the table.

  The updated route information holding unit 49 has topology information in the network, and has its own topology information based on the entry notified from the route optimization processing unit 48 and the subscriber user accommodation node to which the entry should be set and the table. And the function of notifying the traffic information sending / receiving unit 44 of the entry notified from the route optimization processing unit 48 and the subscriber user accommodating node and table to which the entry is set as route information. Yes.

  The load distribution calculation unit 45 performs calculation for assigning a physically different transfer path for each virtual private network using the physical structure of the backbone network and the number of virtual private networks as inputs, and based on this result, The forwarding table of the frame forwarding node is remotely set. Thereby, centralized management type and centralized control type traffic engineering by the server can be realized.

  In this embodiment, a network control server having such a configuration is provided, and the virtual private network technology is applied to the backbone network, and the virtual network is created for each physically different transfer path on the core network that constructs the backbone network. Since the private network is set, it is possible to avoid congestion and failure due to physical causes such as link disconnection and device congestion.

[Operation of Second Embodiment]
Next, the operation of the load distribution type communication network according to the second exemplary embodiment of the present invention will be described with reference to FIGS. FIG. 16 is an explanatory diagram showing an operation of the network control server 5 used in the load distribution type communication network according to the second embodiment of the present invention. FIG. 17 is an explanatory diagram showing a state before frame distribution in the communication network of the operation example. FIG. 18 is an explanatory diagram showing a state after frame distribution in the communication network of the operation example. Here, an example of route change by route optimization processing focusing on the number of transfer frames will be described.

  As shown in FIG. 17, in the state before the frame distribution, in the communication network of the operation example, the route indicated by the route identifier R1 via the frame transfer device 6 is used by the user flow A and the user flow B. Whereas the route is relatively congested, the route indicated by the route identifier R3 via the frame transfer device 8 uses only the user flow E and has a relatively large margin.

  The network control server 5 monitors the transfer load of each device in the communication network, and specifically refers to the monitoring information of the devices in the network periodically. This monitoring information indicates the number of transfer frames between subscriber user accommodation nodes, the total frame length, and the load value observed in each device.

  The periodic reference described above is realized by the traffic information exchanging unit 44. The traffic information exchange unit 44 has a traffic information exchange table 46 shown in FIG. FIG. 19 shows a configuration example of the traffic information exchange table 46, in which timers and output links are managed in association with each destination address. In this operation example, the timer has “600” as an initial value, and the traffic information exchanging unit 44 refers to the monitoring information of each device every 600 seconds.

  The network control server 5 analyzes the monitoring information collected by the traffic information sending / receiving unit 44 in the traffic information holding unit 47 of the load distribution calculating unit 45. When congestion of an arbitrary route is detected, the route optimization processing unit 48 of the load distribution calculation unit 45 performs route optimization in the network.

  When it becomes necessary to change the transfer table of each device in the network due to this route optimization, the change contents are reflected in the topology data of the update route information holding unit 49, and the transfer table entry is added to each device in the network. An SNMP setting request described to change is sent.

  For example, when route optimization is performed by paying attention to the number of transfer frames, congestion in the frame transfer device 6 is detected based on the counter value obtained from the counter table 38 of the subscriber user accommodation node 1, and the user The path is changed so that the path of the frame transfer apparatus 7 in which the flow B is relatively not crowded is used. As a result, as shown in FIG. 18, the forwarding table entry of the subscriber user accommodation node is changed so that the number of forwarding frames of each route becomes equal.

  Next, another operation example of the network control server 5 will be described with reference to FIGS. FIG. 20 is an explanatory diagram showing a failure occurrence state after frame distribution in the communication network of the operation example. FIG. 21 is an explanatory diagram illustrating a state after failure detouring in the communication network of the operation example. Here, an example of route detouring by route optimization processing based on the number of transfer frames and failure detection will be described.

In FIG. 20, a failure has occurred in the path indicated by the path identifier R3 via the frame transfer apparatus 8 in the state after the frame distribution.
The network control server 5 monitors the viability of each device in the communication network. Specifically, the viability of each device can be confirmed by periodically referring to the monitoring information of the devices in the network. . Also, the subscriber user accommodation nodes 1 to 4 regularly check the viability of each route by transmitting and receiving Ping packets between the subscriber user accommodation nodes.

  The subscriber user accommodation nodes 1 to 4 refer to the survivability confirmation table 43 shown in FIG. 22 by the output IF unit 28 and transmit a Ping response request to each subscriber user accommodation node using each transfer route. To do. FIG. 22 is a configuration example of the survivability confirmation table 43, and a timer for periodically transmitting a Ping response request and a survivability confirmation flag are managed in association with each destination core address. ing.

This timer has an initial value of “100” and is counted down by 1 every second. Each time the timer counts down by 20, it sends a Ping response request. The timer becomes 0 (zero) after 100 seconds, but a flag is set when a Ping response is received within 100 seconds.
If the flag is not set even when the timer becomes 0 (zero), it is determined that a failure has occurred in the route, and the network control server 5 is informed that the route indicated by the destination core address has failed. Send the described failure notification.

  The network control server 5 analyzes the monitoring information collected by the traffic information sending / receiving unit 44 in the traffic information holding unit 47 of the load distribution calculating unit 45. And, when a failure of any route is detected, when monitoring information cannot be collected and a failure of any device is detected, or when a notification about a failure of each route is received from the subscriber user accommodation node, The route optimization processing unit 48 of the load distribution calculation unit 45 performs route optimization within the communication network.

  At this time, the failed route and device are excluded from the route assignment target. Because the route optimization needs to change the transfer table of each device in the network, the change contents are reflected in the topology data of the update route information holding unit, and the transfer table entry is changed to each device in the network. The SNMP setting request described as follows is transmitted.

  For example, when optimization is performed focusing on the number of transfer frames, a failure in the frame transfer device 8 is detected based on the inability to collect monitoring information from the frame transfer device 8, and the hot water flow B and the user flow E are Then, the route is detoured so that the routes of the frame transfer device 6 and the frame transfer device 9 which are relatively not crowded are used. As a result, as shown in FIG. 20, the forwarding table entry of the subscriber user accommodation node is changed so that the number of forwarding frames in each route other than the frame forwarding device 8 becomes equal.

  Thus, by having the above-described configuration as the network control server, centralized management type and centralized control type traffic engineering can be realized.

In addition, the virtual private network technology is applied to the backbone network, and the virtual private network is set for each physically different transfer path on the core network that constructs the backbone network. Congestion and failure due to physical causes such as can be avoided.
Thereby, for example, the number of frames transferred on each transfer path in the network per unit time is made uniform, and the total frame length of frames transferred on each transfer path in the network per unit time is made uniform. Alternatively, it is possible to equalize the total load when transferring frames on each transfer path in the network per unit time.

  In the network control server, the physical structure of the backbone network and the number of virtual private networks are input, and a calculation is performed to allocate different physical transfer paths for each virtual private network. Since the forwarding table of the frame forwarding node is remotely set, centralized management type and centralized control type traffic engineering by the server can be realized.

  In addition, the network control server collects the number of transfer frames for each transfer path from the frame transfer nodes in the backbone network, and if there is a difference in the number of transfer frames between the transfer paths, the subscription is made so that this difference is reduced. Since the forwarding table in the user user accommodation node is controlled, the number of frames transferred on each forwarding path in the network per unit time can be made uniform.

  At this time, the network control server collects the total frame length for each transfer path instead of the number of transfer frames, and if there is a difference in the total frame length between the transfer paths, the subscriber user The forwarding table in the accommodating node may be controlled. As a result, the total frame length transferred on each transfer path in the network per unit time can be made uniform.

  Alternatively, the network control server collects the load value consisting of the number of transfer frames and the total frame length for each transfer route, and if there is a difference in the load value between the transfer routes, the subscriber users can be accommodated so that this difference is reduced. The forwarding table in the node may be controlled. As a result, the total load transferred on each transfer path in the network per unit time can be made uniform.

  In addition, the network control server monitors a failure in the backbone network, and when a failure is detected, the transfer path affected by the failure is identified, and the identified information is notified to all subscriber user accommodation nodes. Alternatively, centralized management of the network state by the server can be realized, and as a result, it is possible to detect the failure and specify all the routes affected by the failure.

  Also, when the transfer route related to the failure is notified from the network control server in the subscriber user accommodation node, if the transfer route selected at the time of frame transfer is the same as the notified transfer route, a different transfer route is used. May be reselected, and failure bypassing can be realized in traffic engineering.

  The traffic engineering shown here refers to a technique for equalizing the number of frames transferred on each transfer route in the network per unit time, or on each transfer route in the network per unit time. In some cases, a technique to equalize the total frame length of frames to be transferred, or a technique to equalize the total load when transferring frames on each transfer path in the network per unit time. Is defined based on the policy set in the load observation unit of the subscriber user accommodation node.

For example, when the load monitoring unit enables the threshold of the frame count unit, the network control server equalizes the number of frames transferred on each transfer path in the network per unit time.
Further, in addition to such traffic engineering, centralized management of the network state by the network control server can be realized, and as a result, it becomes possible to identify all routes affected by failure detection and failure.

It is a block diagram which shows the network model of the load distribution type communication network concerning the 1st Embodiment of this invention. It is an example of a physical model of the load distribution type communication network according to the first exemplary embodiment of the present invention. It is an example of a logical model of the load distribution type communication network concerning the 1st embodiment of the present invention. It is a block diagram which shows the structure of the subscriber user accommodation node used with the load distribution type communication network concerning the 1st Embodiment of this invention. It is an example of a physical model of a communication network used in an operation example. It is an example of a logical model of a communication network used in an operation example. It is explanatory drawing which shows operation | movement of a subscriber user accommodation node. It is a structural example of a user frame and a core frame. It is an example of a structure of the destination specific table used by the transfer control part of a subscriber user accommodation node. It is a structural example of the path | route specific table used by the transfer control part of a subscriber user accommodation node. It is a structural example of the counter table used in the load observation part of a subscriber user accommodation node. It is a structural example of the zone | band observation table used in the load observation part of a subscriber user accommodation node. It is a structural example of the load value management table used in the load observation part of a subscriber user accommodation node. It is an example of a structure of the output link specific table used by the output IF part of a subscriber user accommodation node. It is a block diagram which shows the structure of the network control server used with the load distribution type communication network concerning the 2nd Embodiment of this invention. It is explanatory drawing which shows operation | movement of a network control server. It is explanatory drawing which shows the state before the flame | frame dispersion | distribution in the communication network of an operation example. It is explanatory drawing which shows the state after the flame | frame dispersion | distribution in the communication network of an operation example. It is a structural example of the traffic information transfer table used by the traffic information transfer part of a network control server. It is explanatory drawing which shows the failure occurrence state after the flame | frame dispersion | distribution in the communication network of an operation example. It is explanatory drawing which shows the state after a failure detour in the communication network of an operation example. It is a structural example of the survivability confirmation table used by the output IF part of a subscriber user accommodation node.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1-4 ... Subscriber user accommodation node, 5 ... Network control server, 6-9 ... Frame transfer apparatus, 10-17, 50 ... Terminal, 18-21 ... User network, 22 ... Core network, 23 ... User network, 25 ... Input IF unit, 26 ... Transfer control unit, 27 ... Load observation unit, 28 ... Output IF unit, 29 ... Server IF unit, 30 ... Survivability response unit, 31 ... Transfer table, 32 ... Destination specification table, 33 ... Route Specific table 34... Frame count section 35. Total frame length observation section 36. Load value calculation section 37 37 Feedback control section 38. Counter table 39. Band observation table 40 40 Load value management table 41. Output link identification table 42... Survival confirmation unit 43. Survival confirmation table 44. Traffic information transfer unit 45 45 Load distribution calculation unit 46 Traffic information transfer table, 47 ... traffic information holding section, 48 ... route optimization processing unit, 49 ... update path information holding unit, 101 to 124 ... link, 201 ... user frame, 202 ... core frame.

Claims (17)

A backbone network is formed by connecting a plurality of subscriber user accommodation nodes that accommodate a plurality of subscriber users in a network using a plurality of transfer paths each having a unique path identifier. When transferring a frame storing user data in the communication network, the subscriber user accommodation node clearly indicates a transfer route of the frame using the route identifier,
The subscriber user accommodation node is:
A forwarding table for deriving a destination subscriber user accommodation node for the destination user address and a route identifier of the output destination forwarding route;
When a frame received from a subscriber user is transmitted to a subscriber user accommodation node different from itself, an output destination forwarding route is determined based on a route identifier obtained from the forwarding table according to a destination user address of the frame. A transfer control unit;
The number of transfer frames transferred per unit time of each frame having the same path identifier as the destination user address, the destination subscriber user accommodation node, and the number of transfer frames for each pair of the subscriber user accommodation node and the path identifier are totaled. Counter table to manage the total value obtained by
A frame count unit that updates the number of transfer frames and the total value of the corresponding entry in the counter table based on a destination user address, a destination subscriber user accommodation node, and a path identifier extracted from the frame at the time of frame transfer. When,
If the total number of transfer frames for any subscriber user accommodation node / path identifier pair in the counter table exceeds a threshold, one of the route identifiers addressed to the subscriber user accommodation node excluding the path identifier A feedback control unit for selecting from the counter table,
The transfer control unit selects a route selected by the feedback control unit for an arbitrary entry of a destination user address to which the route identifier whose total value exceeds a threshold is assigned in the transfer table. A communication network characterized by assigning an identifier. The communication network according to claim 1,
Collecting the number of frames transferred per unit time for each transfer path from a frame transfer node provided in the backbone network for transferring frames between the transfer paths, and transferring the transfer frames between the transfer paths A communication network, further comprising a network control server that controls to change a forwarding table in the subscriber user accommodation node so that the difference is reduced when the number is different. In the communication network according to claim 1 or 2,
A communication network using a load value obtained by combining the number of transfer frames transferred per unit time and the total frame length transferred per unit time, instead of the number of transfer frames. The communication network according to claim 1,
The feedback control unit, when selecting the route identifier, of the pair of subscriber user accommodation nodes and route identifiers managed in the counter table, the route of the pair whose total number of transfer frames is the smallest A communication network characterized by selecting an identifier. The communication network according to claim 1,
The network further comprises a network control server that monitors a failure in the backbone network and, when a failure is detected, specifies a transfer path affected by the failure and notifies the subscriber user accommodation node network. The communication network according to claim 5 .
In the frame transfer, the transfer control unit selects a transfer route different from the transfer route affected by the failure notified from the network control server as a transfer route used for transferring the frame. network. The communication network according to claim 1,
The backbone network includes a plurality of virtual private networks that connect the subscriber user accommodation nodes, and a physically different transfer path is set for each of the virtual private networks. The communication network according to claim 7 .
Based on the physical structure of the backbone network and the number of virtual private networks, a calculation is performed to allocate a physically different transfer path for each virtual private network, and the calculation result is provided in the backbone network according to the calculation result. A network control server that sets a physically different transfer path for each virtual private network by remotely setting a transfer table of the frame transfer node for a frame transfer node that transfers a frame between the transfer paths; A communication network characterized by comprising. The communication network according to claim 1,
The transmission control unit encapsulates a frame in a predetermined backbone network format when transmitting a user network format frame received from an accommodated subscriber user to the backbone network. A backbone network is formed by connecting a plurality of subscriber user accommodation nodes that accommodate a plurality of subscriber users in a network using a plurality of transfer paths each having a unique path identifier. A subscriber user accommodation node used in a communication network for clearly indicating a transfer route of the frame using the route identifier in the subscriber user accommodation node when transferring a frame storing user data in
A forwarding table for deriving a destination subscriber user accommodation node for the destination user address and a route identifier of the output destination forwarding route;
When a frame received from a subscriber user is transmitted to a subscriber user accommodation node different from itself, an output destination forwarding route is determined based on a route identifier obtained from the forwarding table according to a destination user address of the frame. A transfer control unit;
The number of transfer frames transferred per unit time of each frame having the same path identifier as the destination user address, the destination subscriber user accommodation node, and the number of transfer frames for each pair of the subscriber user accommodation node and the path identifier are totaled. Counter table to manage the total value obtained by
A frame count unit that updates the number of transfer frames and the total value of the corresponding entry in the counter table based on a destination user address, a destination subscriber user accommodation node, and a path identifier extracted from the frame at the time of frame transfer. When,
If the total number of transfer frames for any subscriber user accommodation node / path identifier pair in the counter table exceeds a threshold, one of the route identifiers addressed to the subscriber user accommodation node excluding the path identifier A feedback control unit for selecting from the counter table,
The transfer control unit selects a route selected by the feedback control unit for an arbitrary entry of a destination user address to which the route identifier whose total value exceeds a threshold is assigned in the transfer table. A subscriber user accommodation node characterized by assigning an identifier. The subscriber user accommodation node according to claim 10 ,
A subscriber user accommodation node using a load value obtained by combining the number of transfer frames transferred per unit time and the total frame length transferred per unit time instead of the number of transfer frames. The subscriber user accommodation node according to claim 10 ,
The feedback control unit, when selecting the route identifier, of the pair of subscriber user accommodation nodes and route identifiers managed in the counter table, the route of the pair whose total number of transfer frames is the smallest A subscriber user accommodation node, wherein an identifier is selected. The subscriber user accommodation node according to claim 10 ,
The transfer control unit encapsulates the frame in a predetermined backbone network format when transmitting the user network format frame received from the accommodated subscriber user to the backbone network. Containment node. A backbone network is formed by connecting a plurality of subscriber user accommodation nodes that accommodate a plurality of subscriber users in a network using a plurality of transfer paths each having a unique path identifier. When a frame storing user data is transferred, the subscriber user accommodation node uses the path identifier to specify the transfer path of the frame, and controls the path selection at the subscriber user node. A network control server that
A traffic information transfer unit that collects the number of frames transferred per unit time for each transfer path from a frame transfer node that is provided in the backbone network and transfers frames between the transfer paths;
A network control server comprising: a load distribution calculation unit that controls and changes a forwarding table in the subscriber user accommodation node so that the difference is reduced when there is a difference in the number of transfer frames between the transfer paths. . The network control server according to claim 14 ,
A network control server using a load value obtained by combining the number of transfer frames transferred per unit time and the total frame length transferred per unit time instead of the number of transfer frames. The network control server according to claim 14 ,
The load distribution calculating unit monitors a failure in the backbone network, and when a failure is detected, specifies a transfer path affected by the failure and notifies the subscriber user accommodation node of the failure. Network control server. The network control server according to claim 14 ,
The load distribution calculation unit performs a calculation to allocate a physically different transfer path for each virtual private network based on the physical structure of the backbone network and the number of the virtual private networks, and according to the calculation result For a frame forwarding node that is provided in a backbone network and forwards frames between the forwarding routes, a forwarding table that the frame forwarding node has is remotely set to set a physically different forwarding route for each virtual private network. A network control server characterized by setting.

Images