JP2000181823A - Fault tolerance network management system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable a subsidiary sub-manager to economically realize vicarious execution management in the case of fault in a main sub-manager with a simple configuration and secure fault tolerance performance to the fault of the main sub-manager. SOLUTION: This system is provided with plural local area networks 1, 2 and 3 connected to a wide area network 5 and main sub-managers 11 and 12 which are connected to the local area networks 1, 2 and 3 having no redundant configuration. The plural subsidiary sub-managers 14 and 15 having the redundant configuration being smaller than that in the main sub-managers at the side of an integrated manager is allowed to vicariously execute the function of either one of the main sub-managers 11 and 12 where fault occurs.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to an agent,
The present invention relates to a fault-tolerant network management system that is hierarchically managed by a sub-manager and an integrated manager.

[0002]

2. Description of the Related Art Conventionally, in a fault-tolerant network management system, when a catastrophic failure occurs in a system, in order to perform automatic restoration without interrupting service, as disclosed in Japanese Patent Application Laid-Open No. Hei 9-116538, for example. In addition, one sub-network management system is always provided for one of the main network management systems, thereby ensuring fault tolerance. Also,
When managing a distributed network such as a large-scale network or a local area network (hereinafter, referred to as a LAN), a hierarchical structure of agents, sub-managers, and integrated managers is used to reduce management packets and simplify the network management system. In some cases, it is more effective to manage a network by arranging a network. In this case, as shown in the above-mentioned publication, a configuration in which one main network management system and one sub-network management system are always used Then the main sub-manager m
When a number of sub-sub-managers are arranged and a hierarchical structure management for performing network management is adopted, m sub-sub-managers also need to be arranged.

[0003]

However, in such a conventional fault-tolerant network management system, the main network management system and the sub-network management system have a fixed one-to-one system configuration. When there are m sub-network management systems, m sub-network management systems are necessarily required. As a result, not only the configuration becomes complicated, but also the system becomes expensive. Further, in order to ensure fault tolerance of the network management system, even if there are m main network management systems and m sub-network management systems, if a pair of main / sub-network management systems suffers a double failure. However, there has been a problem that fault-tolerance cannot be ensured.

SUMMARY OF THE INVENTION The present invention solves the above-mentioned problems. A plurality of main sub-managers are provided with a smaller number of sub-sub-managers. It is an object of the present invention to provide a fault-tolerant network management system capable of ensuring fault-tolerance against a failure of a main sub-manager while realizing the above.

[0005]

To achieve the above object,
The fault tolerant network management system according to the first aspect of the present invention includes a plurality of local area networks connected to a wide area network via a router, and a plurality of local area networks connected to some of these local area networks. A main sub-manager that does not have a redundant configuration for managing and controlling managed objects in the local area network is provided to an integrated manager having a redundant configuration connected to the rest of the local area network via the main sub-managers and the wide area network. One of the sub-sub-managers, which is smaller than the total number of the main sub-managers connected to the local area network to which the integrated manager is connected, manages and controls these managed objects. It is obtained so as to act for the function of the main sub-manager.

In the fault tolerant network management system according to the second aspect of the present invention, the number of the sub-sub-managers is set on the basis of the probability with which the main sub-manager is likely to fail simultaneously. Things.

In the fault tolerant network management system according to the third aspect of the present invention, an agent that manages and controls a management object such as configuration information and status information for each resource unit is connected to each of the local area networks. Things.

The fault tolerant network management system according to the invention of claim 4 is characterized in that each of the sub-sub-managers and the corresponding agent, between the integrated manager and each of the main sub-managers, and between each of the sub-sub-managers. The management object is managed between the manager and each of the agents and between the integrated manager and each of the sub-sub-managers using a simple network management protocol.

[0009]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of the fault tolerant network management system of the present invention. In the figure, 1, 2, and 3 indicate a plurality of LANs.
These are wide area networks (hereinafter referred to as WAN).
5) are connected via routers 6, 7, and 8. The LAN 2 is connected to a plurality of agents 20 and 21 for managing and controlling management objects such as configuration information and status information in units of network resources.
A main sub-manager 11 having no redundant configuration for managing and controlling the management objects in the LAN 2 is connected.

In the same manner as described above, the LAN 3 has a plurality of agents 30 and 3 for managing and controlling management objects such as configuration information and status information in network configuration units.
The main sub-manager 12 having no redundant configuration for managing and controlling the management objects in the LAN 3 is connected via these agents 30 and 31. On the other hand, the LAN 1 has a plurality of agents 5
0, 51 are connected, and these agents 50, 51
An integrated manager 10 that manages and controls the managed objects under the management of 51, and manages and controls these managed objects through the WAN 5 and the main sub-managers 11 and 12 is connected. That is, L
An integrated manager 10 that manages resources of the entire network is connected to AN1.

The LANs 2 and 3 are sites managed by the integrated manager 10, and the total number is m. That is, the integrated manager 10 manages and controls each of the managed objects under the management through the total number m of the main sub-managers 11 to 12. Here, only the integration manager 10 itself has a one-to-one redundant configuration. Further, the LAN 1 is connected to sub-sub-managers 14 and 15 which are to be backups of the m main sub-managers 11 and 12 of the total number of LANs 2 and 3. The sub-sub-managers 14 and 15 are backup sub-managers that take over the function of the main sub-manager when any one of the main sub-managers 11 and 12 has a catastrophic failure. Incidentally, there are a total of n sub-sub-managers 14 and 15. At this time, m> n, and the value of n is the main sub-manager 11, 1
2 is determined based on the probability of simultaneous failure.

FIG. 2 shows agents 20, 21, 30,
FIG. 3 is an explanatory diagram showing a logical relationship among 31, main sub-managers 11 and 12, sub-sub-managers 14 and 15, and an integrated manager 10. Here, between the main sub-manager 11 and the agents 20 and 21, an international standard IAB (Internet Ac) for communication network management.
SNMP (Simp)
le Network Management Pro
tocol) to manage managed objects.
Then, the main sub-manager 11 is a MIB (Managem) that expresses a set of a plurality of management objects collected through the agents 20 and 21 in the management range in a tree structure.
entInformation Base).

Similarly, between the main sub-manager 12 and the agents 30 and 31, managed objects are managed by using SNMP.
12 sub-sub-managers 14 and 15 to be backup
And between each agent 20, 21, 30, 31
Manage managed objects using NMP. In addition,
Between the main sub-managers 11 and 12 and the integrated manager 10, the sub-sub-managers 14 and 15 and the integrated manager 10
Also, the management object is managed using SNMP.

FIG. 3 and FIG.
1, main sub-managers 11 and 12, sub-sub-manager 1
4, 15 When the network resources are hierarchically managed by the integrated manager 10, the main sub-managers 11, 12
2 shows a sequence of operations of each system for ensuring fault-tolerance as a network management system when the secondary sub-managers 14 and 15 to be backed up are activated. The integrated manager 10 sends ICMP (Internet Control) to the m main sub-managers of the main sub-managers 11 and 12.
ol Message Protocol), an ICMP echo request is issued periodically (health check), and the resulting ICMP echo response is obtained, thereby confirming whether each of the main sub-managers 11 and 12 is alive or dead.

As described above, for example, when there is no response from the main sub-manager 12 even after a certain period of time has elapsed since the integrated manager 10 issued the ICMP echo request to each of the main sub-managers 11 and 12 (response waiting timeout) ), And recognizes that the main sub-manager 12 has failed. Then, the integrated manager 10 becomes the main sub-managers 11 and 12.
When the failure of the main sub-manager 12 is recognized,
The unified manager 10 replaces the free sub-manager 14 with the failed main sub-manager 1, for example.
A message requesting the proxy of the second is transmitted using SNMP. This request message is sent together with the IP address of the main sub-manager 12 to be substituted, the network address of the LAN 3 to which the sub-manager 12 belongs, and the address of the router connecting the LAN 3. The integrated manager 10 manages the in-use state of the sub-sub-managers 14 and 15, and the information includes a used / unused state, and when in use, indicates which main sub-manager is acting on behalf of Manages IP address information.

The sub-sub-manager 14 that has received the proxy request from the main sub-manager 12 includes the LA to which the main sub-manager 12 to be substituted belongs, which is included in the proxy request message.
Based on the IP address of the router 8 connecting N3, SN
RFC 1 that the router has by MP
213, Management Information Base for Network Management of TSP / IP Based Internet, Mib-two (RFC1213, Management In
formation Base for Network
k Management of TCP / IP Ba
dot notation defined by the sed Internet (MIB-II), iso (1). ord (3). dod
(6). internet (1). mgmt (2). m
ib-2 (1). at (3). Acquires all the values of atTable existing in atTable (1).

The value of the above atTable is determined by the router 8 and the ARP (address Resolution).
Protocol), the physical address is resolved from the IP address, and the correspondence between all the IP addresses and the physical addresses is retained. As a result, the device connected to the LAN 3 accommodated by the router 8, that is, the LAN 3 It is possible to acquire all combinations of IP addresses and physical addresses of the connected agents 30 and 31.

As the next step, the sub-sub-manager 1
4 is for all the agents 30 and 31 connected to the LAN 3 based on the IP address obtained above.
A message for changing the IP address of the own sub-sub-manager 14 is transmitted by SNMP, and the agents 30 and 31
Is notified that the manager has changed from the main sub-manager 12 to the sub-sub-manager 14. As described above, the sub-sub-manager 14 manages the I / Os of all the agents 30 and 31 to manage and control the configuration information, the status information, and the like.
The agents 30 and 31 that have acquired the P address and the main sub-manager 12 has failed can acquire the IP address of the new substitute manager, that is, the sub-sub-manager 14.

Thereafter, the integrated manager 10 transmits various messages that have been transmitted to the failed main sub-manager 12 to the sub-sub-manager 14. The sub-sub-manager 14 transmits the message to each of the agents 30 and 31 under the corresponding LAN 3, and traps destined for the sub-sub-manager 14 are also sent as needed.
By transmitting to the network management system, the fault tolerance of the entire network management system is ensured. Since the integrated manager 10 recognizes that the number of active sub-sub-managers 14 to be managed has increased, the integrated manager 10 also periodically sends an ICMP echo request to the newly activated sub-sub-managers 14, As in the case of the main sub-manager 12, the life and death are monitored.

Therefore, the sub-submanager 14 to be monitored
Does not respond to the ICMP echo request from the integrated manager 10 even after a certain period of time, the integrated manager 10 transmits a proxy request to the next free sub-sub-manager 15 in the same procedure as described above, and The sub-manager 14 is executed. As described above, as long as a plurality of main sub-managers 11 and 12 do not fail simultaneously, if there are n sub-sub-managers 14 and 15, when focusing on one certain main sub-manager, fault tolerance up to n-fold failures Can be ensured. On the other hand, the integrated manager 10
Also, the main sub-manager 12 continues to periodically transmit the ICMP echo request and monitors the recovery of the main sub-manager 12. At a certain timing, the main sub-manager 12 that has
When a response to the MP echo request is transmitted, the integrated manager 10 determines that the main sub-manager 12 that has been in failure so far has been recovered.

At this time, the integrated manager 10 sends a restoration message to the restored main sub-manager 12 by SNMP, and the sub-
0, 31 is instructed to transmit a message from the sub-sub-manager 15 to the main sub-manager 12 indicating that the manager IP address has been changed. By this message, each agent under the LAN recognizes that the main sub-manager has been restored, and each system is restored to the original state. At the same time, the integrated manager 10 transmits a message of delegation release to the sub-sub-manager 15 that has been acting for the main sub-manager 12, manages the state of the sub-sub-manager managed by the integrated manager 10 to be empty, and Prepare for a failure of any of the sub-managers 11 and 12.

[0022]

As described above, according to the present invention, the system configuration is such that the main network management system and the sub network management system are not fixedly one-to-one, so that the faults of the m network management systems In order to ensure the tolerant property, the effect that the fault tolerant property of the network management system can be ensured by n sub-sub-managers having a value smaller than m and that this can be realized with a simple and low-cost configuration is obtained. Further, as long as a plurality of main sub-managers do not fail at the same time, the effect that fault tolerability can be ensured up to a maximum n-fold failure of one main sub-manager is obtained.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a fault tolerant network management system according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing a logical relationship among an agent, a sub-manager, and an integrated manager in FIG.

FIG. 3 is a sequence diagram showing operations of an agent, a sub-manager, and an integrated manager in FIG.

FIG. 4 is a sequence diagram showing operations of the agent, the sub-manager, and the integration manager in FIG. 1, following FIG. 3;

[Explanation of symbols]

 1,2,3 LAN (Local Area Network) 5 WAN (Wide Area Network) 6,7,8 Router 10 Integrated Manager 11,12 Main Sub-Manager 14,15 Sub-Sub Manager 20,21,30,31,50,51 Agent

────────────────────────────────────────────────── ───

[Procedure amendment]

[Submission date] March 16, 2000 (200.3.1.1)
6)

[Procedure amendment 1]

[Document name to be amended] Statement

[Correction target item name] Claims

[Correction method] Change

[Correction contents]

[Claims]

[Procedure amendment 2]

[Document name to be amended] Statement

[Correction target item name] 0005

[Correction method] Change

[Correction contents]

[0005]

To achieve the above object,
The fault tolerant network management system according to the first aspect of the present invention includes a plurality of local area networks connected to a wide area network via a router, and a plurality of local area networks connected to some of these local area networks. A main sub-manager that does not have a redundant configuration for managing and controlling managed objects in the local area network is provided to an integrated manager having a redundant configuration connected to the rest of the local area network via the main sub-managers and the wide area network. One of the sub-sub-managers, which is smaller than the total number of the main sub-managers connected to the local area network to which the integrated manager is connected, manages and controls these managed objects. It has a configuration which act for the function of the main sub-manager. And the integration manager
If the master recognizes the failure of the primary sub-manager,
The manager issues an error to the free secondary sub-manager.
Message requesting harmful primary sub-manager to act
Is transmitted using SNMP, and the request message is transmitted.
Contains the IP address of the main sub-manager to be substituted,
Network address of the LAN to which the
The next router address is transmitted together .

Claims (4)

[Claims] 1. A plurality of local area networks connected to a wide area network via a router, and connected to some of these local area networks to manage and control a management object in each of these local area networks. A main sub-manager having no redundant configuration; an integrated manager having a redundant configuration connected to the rest of the local area network and managing and controlling these managed objects via the main sub-managers and the wide area network; A plurality of sub-sub-managers connected to the local area network to which the integrated manager is connected and substituting for the function of any of the failed main sub-managers, the number being smaller than the total number of the main sub-managers. Fall Tolerant network management system. 2. The fault according to claim 1, wherein the number of the sub-sub-managers is set based on the probability that the main sub-manager is likely to fail simultaneously. Tolerant network management system. 3. The local area network according to claim 1, wherein an agent that manages and controls a management object such as configuration information and status information for each resource unit is connected to each of the local area networks. Fault tolerant network management system. 4. Each of the main sub-managers and their corresponding agents, between the integration manager and each of the main sub-managers, between each of the sub-sub-managers and each of the agents, and between the integration manager and each of the agents. The fault tolerant network management system according to claim 1, wherein the management object is managed between the sub-manager and the sub-manager using a simple network management protocol.