JP3803233B2 - Network error handling method and node device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a method for coping with a case where the entire network falls into an abnormal state in a network adopting a dynamic routing method, and a node device therefor.
[0002]
[Prior art]
Create and update the network topology database for creating routing tables (route tables) in OSPF (Open Shertest Path First) in IP (Internet Protocol) networks and PNNI (Private Network-Network Interface) in ATM (Asynchronous Transfer Mode) networks Therefore, a dynamic routing method using a link state algorithm is employed.
[0003]
[Problems to be solved by the invention]
This method automatically updates the topology database of the network held by each node by sending its link state to the network periodically and when each node changes its configuration. There are advantages such as "and routing loop is hard to occur". However, due to the principle of operation, if a node in the network malfunctions and an incorrect link state flows to the network, it affects the topology database of the entire network, and the network routing operation is paralyzed. Can happen.
[0004]
Accordingly, an object of the present invention is to provide a coping method when the entire network falls into an abnormal state in a network adopting a dynamic routing method and a node device therefor.
[0005]
[Means for Solving the Problems]
According to the present invention, by exchanging the link state of each node among a plurality of nodes included in the network, each node maintains the same topology data and uses it for routing control. As a countermeasure, snapshots of topology data when the network is in a stable state are recorded at the same time in each node, and the snapshots are used as topology data in each node while the network is in an abnormal state. A method is provided comprising the steps used for routing control.
[0006]
According to the present invention, by exchanging the link state of each node among a plurality of nodes included in the network, each node maintains the same topology data and copes with a network abnormality in a network used for routing control. And a means for recording a snapshot of topology data when the network is in a stable state at the same time as other nodes, and the snapshot while the network is in an abnormal state. And a node device comprising means for use in routing control.
[0007]
DETAILED DESCRIPTION OF THE INVENTION
Prior to the description of the present invention, a dynamic routing method in OSPF in an IP network or PNNI in an ATM network will be described.
In the network configuration shown in FIG. 1, the network A is connected to the network C via the node 2. Here, a router corresponds to a node in the IP network, and an ATM switch corresponds to a node in the ATM network. The network C is further connected to the network D via the node 1. More specifically, network C is directly connected to port 1 of node 1 and network D is directly connected to port 0 of node 1.
[0008]
Each node constructs a topology database storing topology data as shown in FIG. 2 by exchanging link states indicating which networks are directly connected to each node according to a link state algorithm. Based on this topology data, each node creates a routing table (route table), and IP packets or ATM cells are routed in accordance with the routing table. FIG. 3 shows a routing table of the node 1 in the example of FIG. According to the routing table of FIG. 3, for example, an IP packet or ATM cell whose destination network is A that has reached node 1 is routed from port 1 to node 2.
[0009]
The packet or cell includes a general user data packet or cell and a routing protocol packet or cell. The user data packet or cell is routed according to the routing table as described above. In the case of a packet or cell for a routing protocol, the operation varies depending on the type, but in the case of a packet or cell carrying the above link state, it is transferred to another node and stored in the topology database as the topology data shown in FIG. Based on this, a new routing table is created.
[0010]
FIG. 4 shows an example of a network configuration to which the present invention is applied. In FIG. 4, the nodes 10 are connected to each other via a network 12 and are also connected to an NMS (Network Management System) 14 for managing the entire network.
FIG. 5 shows a configuration of the node device 10 according to an embodiment of the present invention. The node device corresponds to a router in an IP network, for example, and corresponds to an ATM switch in an ATM network. In FIG. 5, the NMS control unit 20 exchanges data with the NMS in FIG. 4, and the port control unit 22 routes packets or cells input / output from the port 24 according to the routing table 25. The port control unit 22 corresponds to an ATM switch in the case of an ATM exchange. When the arriving packet or cell is a routing protocol packet or cell, the packet or cell is also sent to the main processor 26. If the link state is included in the routing protocol packet or cell, the topology data 28 is updated accordingly, and a new routing table 25 is created based on the updated topology data 28. The snapshot 30 will be described later.
[0011]
In the embodiment described below, it is assumed that the time of all nodes under the NMS, including the NMS, are completely the same, so if they do not match (not shown here) Keep them in the same way. This time synchronization check is performed at an appropriate cycle, and the time is always matched between the nodes.
FIG. 6 is a flowchart of snapshot collection processing in the main processor 26 of the node device 10. When the NMS 14 instructs the maintenance person to take a snapshot of the network, the NMS 14 designates a time for each subordinate node, and when this time comes, a copy of the topology data held by each node is taken. Prompt. When the time indicated by the NMS 14 is reached (step 1000), each node 10 takes a copy of the topology data (step 1002), and at the same time, verifies whether the data is sufficiently stable from the previous exchange of topology data by the following method. (Step 1004), and the result is notified to the NMS (Step 1008).
[0012]
In general, the packets used in the link state algorithm include a database description packet that carries network configuration information and a Hello packet that indicates that the local node can be used. The latter Hello packet always flows on the network link at a certain interval when the node is operating, but the former database description packet has a change in the network configuration, and synchronization between each node Used only when matching. Therefore, if only Hello packets are flowing on the link of the network for a certain period of time, the network can be regarded as stable (synchronized) at that time.
[0013]
FIG. 7 is a flowchart of stable state determination processing. In FIG. 7, when a packet is received (step 1100), it is determined whether or not it is a hello packet (step 1102). If it is not a hello packet, a stability timer is initialized (step 1104), Stable ”(step 1106). When the received packet is a Hello packet, if the timer has expired (step 1108), the state is set to "stable" (step 1110).
[0014]
It should be noted that the time until this can be regarded as stable (grace time) is the time possessed by each node or the time instructed by the NMS, and each node selects one of these two methods. May be.
The NMS 14 waits for notifications from all the nodes and confirms whether or not sufficiently stable data has been obtained at all the nodes. Through this series of operations, a copy (snapshot) of topology data at a certain time can be held in all nodes in the network.
[0015]
Note that snapshots may be collected autonomously at a time when all nodes are determined without receiving an instruction from the NMS 14 when creating a snapshot. In this case, the processing of other parts is the same as described above.
Thereafter, each node regularly checks the normality of the network according to the procedure shown in FIG. Each node calculates values such as a call loss rate for connection-type connections and a packet discard rate for connectionless connections (step 1202). However, in this calculation, the fault caused by the hardware installed in the machine is subtracted, and only the part generated as a result of routing by software is calculated. When a specific route becomes unusable due to a hardware failure, the routing protocol itself is automatically excluded from the routing target and does not cause call loss or packet discard. When the calculated value exceeds a certain reference value (step 1204), it is determined that the network is in an abnormal state, and a notification to that effect is sent to the NMS (step 1206).
[0016]
This reference value may be a value that is unique to each node or a value that is instructed by the NMS, and each node may select one of these two methods.
FIG. 9 shows NMS processing when a notification of an abnormal state is received from any node. The NMS that has received the abnormal state instructs each node to discard the topology data that is currently used when the instructed time is reached, and to use the topology data that is stored as a snapshot (step 1302). . Each node that receives an instruction from the NMS discards the topology data that has become abnormal data at the specified time, regenerates the routing table from the retained topology data, and uses this data for call processing operations Continue. The cause of the abnormality during this period is identified and recovered, and after topology data that is considered to be normal is completed by the link state algorithm, the NMS is notified (step 1304). In response to this, the NMS again notifies each node of the return time to the normal mode (step 1306), and each node shifts to the normal mode according to the instruction. FIG. 10 shows an operation sequence of the NMS and the node when an abnormality is detected.
[0017]
(Supplementary note 1) A method for dealing with network abnormality in a network in which each node holds the same topology data and is used for routing control by exchanging link states of the nodes among a plurality of nodes included in the network. Because
(A) Record snapshots of topology data when the network is in a stable state at each node at the same time,
(B) A method comprising the step of using the snapshot as topology data for routing control in each node while the network is in an abnormal state. (1)
(Appendix 2) Step (a)
(I) determine the stability of the network at each node;
(Ii) Create snapshots at the same time on each node,
(Iii) The method according to supplementary note 1, including a sub-step of notifying each network management system (NMS) of the fact that a snapshot has been successfully created when it is determined that the network is stable. (2)
(Supplementary Note 3) Sub-steps (a) and (ii)
Instructing each node to create a snapshot at a specified time from the NMS,
The method according to appendix 2, including a sub-step of copying the topology data as a snapshot at a specified time in accordance with an instruction from the NMS at each node.
[0018]
(Supplementary note 4) The method according to supplementary note 2, wherein in each of the sub-steps (a) and (ii), each node copies topology data as a snapshot at a predetermined time regardless of an instruction from the NMS.
(Supplementary note 5) The method according to supplementary note 2, wherein, in sub-steps (a) and (i), it is determined that the network is stable when a link state is not received from an adjacent node within a predetermined grace period.
[0019]
(Supplementary note 6) The method according to supplementary note 5, wherein the grace period is uniquely set for each node.
(Supplementary note 7) The method according to supplementary note 5, wherein the grace time is specified by the NMS.
(Supplementary Note 8) Step (b)
(I) In each node, determine the normality of the network,
(Ii) When it is determined that the network is abnormal, the node notifies the NMS to that effect,
(Iii) The method according to appendix 1, including a sub-step for instructing each node to switch the topology data to a recorded snapshot at a specified time in response to a notification that the network is abnormal. (3)
(Supplementary note 9) The method according to supplementary note 8, wherein, in the substeps (b) and (i), when the call loss rate and / or the packet discard rate excluding those due to hardware failure exceeds a predetermined threshold, it is determined that the network is abnormal.
[0020]
(Supplementary note 10) The method according to supplementary note 9, wherein the threshold value is set uniquely for each node.
(Additional remark 11) The said threshold value is the method of Additional remark 9 designated from NMS.
(Additional remark 12) In order to cope with the network abnormality in the network where each node holds the same topology data and is used for routing control by exchanging the link state of each node among a plurality of nodes included in the network. Node equipment,
Means for recording a snapshot of topology data at the same time as other nodes when the network is in a stable state;
A node device comprising: means for using the snapshot as topology data for routing control while the network is in an abnormal state. (4)
(Supplementary note 13) Snapshot recording means
A means of determining network stability;
Means to create a snapshot at the same time as other nodes;
The node device according to appendix 12, further comprising means for notifying the NMS of the successful creation of a snapshot when it is determined that the network is stable. (5)
(Supplementary note 14) Snapshot creation means
Means for receiving an instruction to create a snapshot at a specified time from the NMS;
14. A node device according to appendix 13, comprising means for copying topology data as a snapshot at a specified time in response to an instruction from the NMS.
[0021]
(Supplementary note 15) The node device according to supplementary note 13, wherein the snapshot creation means copies the topology data as a snapshot at a predetermined time regardless of an instruction from the NMS.
(Supplementary note 16) The node device according to supplementary note 13, wherein the network stability determination means determines that the network is stable when a link state is not received from an adjacent node within a predetermined grace period.
[0022]
(Supplementary note 17) The node device according to supplementary note 16, wherein the grace period is uniquely set for each node.
(Supplementary note 18) The node device according to supplementary note 16, wherein the grace time is designated by the NMS.
(Supplementary note 19) Snapshot use means
Means for determining the normality of the network;
Means for notifying the NMS when the network is determined to be abnormal;
The node device according to appendix 12, including means for switching topology data to a recorded snapshot at a specified time in response to an instruction from the NMS.
[0023]
(Supplementary note 20) The node device according to supplementary note 19, wherein the network normality determining means determines that the network is abnormal when the call loss rate and / or the packet discard rate excluding those due to hardware failure exceeds a predetermined threshold.
(Supplementary note 21) The node device according to supplementary note 20, wherein the threshold value is uniquely set for each node.
[0024]
(Additional remark 22) The said threshold value is a node apparatus of Additional remark 20 designated from NMS.
[0025]
【The invention's effect】
As described above, according to the present invention, it is possible to prevent the worst situation in which the entire network is paralyzed due to an abnormality of one node and the service cannot be provided. Can continue and maintain availability.
[Brief description of the drawings]
FIG. 1 is a diagram illustrating an example of a network configuration.
FIG. 2 is a diagram showing topology data in the network configuration of FIG. 1;
FIG. 3 is a diagram showing a routing table of node 1 generated from the topology data of FIG. 2;
FIG. 4 is a diagram showing an example of a network configuration to which the present invention is applied.
FIG. 5 is a block diagram illustrating a hardware configuration of a node device.
FIG. 6 is a flowchart of snapshot collection processing;
FIG. 7 is a flowchart of a stable state determination process.
FIG. 8 is a flowchart of a node operation normality confirmation process.
FIG. 9 is a flowchart of an operation at the time of abnormality in the NMS.
FIG. 10 is a sequence diagram showing the operation of the NMS and each node at the time of abnormality.

Claims (5)

By exchanging the link state of each node among a plurality of nodes included in the network, each node retains the same topology data and uses it for routing control.
(A) A snapshot of topology data when the network is in a stable state in which only packets other than packets that carry network configuration information flow as packets used in the link state algorithm within a certain time on the link of the network. Record at the same time in the node,
(B) A method comprising the step of using the snapshot as topology data at each node for routing control while the network is in an abnormal state until an abnormality occurs in the network and the network is recovered . By exchanging the link state of each node among a plurality of nodes included in the network, each node retains the same topology data and uses it for routing control.
(A) Record snapshots of topology data when the network is in a stable state at each node at the same time,
(B) including a step of using the snapshot as topology data for routing control in each node while the network is in an abnormal state;
Step (a)
(I) determine the stability of the network at each node;
(Ii) Create snapshots at the same time on each node,
(Iii) when the network is successfully created snapshot when it is determined to be stable, including METHODS a substep notifies each node in the network management system (NMS). Step (b)
(I) In each node, determine the normality of the network,
(Ii) When it is determined that the network is abnormal, the node notifies the NMS to that effect,
(Iii) The method according to claim 1, further comprising a sub-step for instructing each node to switch the topology data to a recorded snapshot at a specified time in response to a notification that the network is abnormal. . A node device for dealing with network anomalies in a network used for routing control by maintaining the same topology data for each node by exchanging the link state of each node among a plurality of nodes included in the network. There,
A snapshot of topology data when the network is in a stable state where only packets other than those that carry network configuration information flow as packets used in the link state algorithm within a certain period of time on the link of the network. Means for recording at the same time;
A node device comprising means for using the snapshot as topology data for routing control while the network is in an abnormal state until an abnormality occurs in the network and the network is recovered . A node device for dealing with network anomalies in a network used for routing control by maintaining the same topology data for each node by exchanging the link state of each node among a plurality of nodes included in the network. There,
Means for recording a snapshot of topology data at the same time as other nodes when the network is in a stable state;
Means for using the snapshot as topology data for routing control while the network is in an abnormal state;
Snapshot recording means
A means of determining network stability;
Means to create a snapshot at the same time as other nodes;
When the network has succeeded in creating a snapshot of when it is determined to be stable, including node device and means for notifying the NMS.

Images