JPH11177552A - Trap management method and system for snmp - Google Patents

Abstract

PROBLEM TO BE SOLVED: To allow a manager to surely obtain a TRAP. SOLUTION: A manager 11 has an SNMP processing section 11-2 that processes an SNMP protocol, a main processing section 11-2 that processes monitor control, and a memory 11-3 that stores a control program and data including a version number of a preceding TRAP received from an agent. The agent 12 is provided with an SNMP processing section 12-1 that processes the SNMP, a main processing section 12-2 that processes monitor control, a memory 12-3 that stores a control program and data including the version number of the TRAP sent to the manager precedingly, and a node communication section 12-4 that makes communication with the node 13. The agent 12 sends the TRAP to which two version numbers being this time and preceding time are added to the manager 11. The manager 11 compares the version number of the TRAP with the version number stored in its own memory, updates the storage of the version number when they are coincident, and requests re- transmission of the TRAP to the agent 12 in the case of dissidence.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

TECHNICAL FIELD The present invention relates to a plurality of nodes,
An agent for monitoring and controlling the node, and the agent are connected to a simple network management protocol (Simp) which is a standard management protocol for a TCP / IP network.
le Network Management Protocol (hereinafter referred to as SNMP)
The present invention relates to an NMP TRAP management method and device.

[0002]

2. Description of the Related Art The SNMP TRAP management system comprises a plurality of nodes, a manager for monitoring and controlling these nodes, and an agent capable of performing monitoring and control of a node to which the manager belongs under remote monitoring and control by the manager. In this network, transmission and reception of TRAP for notifying an event that has occurred in each node to a manager serving as a management station is managed.

In a management system using SNMP, monitoring is generally performed by polling from a management station to a managed node. However, when the number of managed nodes increases, the network load due to polling is increased. And the overhead of the management station increases. Therefore, in order to manage the node in more real time, monitoring is performed on the managed node, and the TR is sent to the management station only when an abnormality occurs.
Transmitting an AP is performed.

However, simply using the SNMP TRAP does not guarantee that the TRAP will reach the other party. Therefore, the “network management system” disclosed in Japanese Patent Laid-Open No. 8-331206 discloses a network management system using SNMP. For the purpose of ensuring the transmission reliability of TRAP transmitted from the managed device to the management device without increasing the memory capacity of the network component device, the management device is provided with a TRAP response processing unit, and the managed device is provided with TRAP response processing.
Each of the response confirmation units is provided, and when the management device receives TRAP transmitted from the managed device, the management device transmits a PDU (Protoclol Data Unit) relating to the specific managed object to the managed device, and the managed device By confirming whether a PDU has been received from the management device within a predetermined period after the TRAP transmission, it is determined whether the transmission of the TRAP has succeeded or failed.

[0005] Further, a "network management system" disclosed in Japanese Patent Application Laid-Open No. Hei 9-64871 distributes management data to a plurality of managers in order to reduce the load on managers that increases with the increase in services. In a network management system in which a manager shares and manages, an agent on a managed device sends an SN to a manager.
Data is sent using MP TRAP. in this case,
Since TRAP that needs to be transmitted to only some of the managers is sent to all the managers, a problem arises in that the processing load on each manager increases, and the object is to solve the problem. The solution is to provide the manager with a TRAP setting request unit for setting the content and destination in the TRAP packet for the agent, and the agent with the content of the TRAP packet included in the packet received by the set-request receiving unit. And a TRAP setting unit for detecting and setting a destination, and a TRAP transmitting unit for transmitting a TRAP packet to the set destination.

[0006] In addition, TRAP is performed using SNMP standard management.
The aim of ensuring that the data is delivered to the management station is disclosed in Japanese Patent Application Laid-Open No. Hei 9-101929, entitled "TRAP transmitter".

That is, when a management station receives a TRAP, a message indicating the normal reception of the TRAP is transmitted to the TRAP transmission source using SNMP, and the managed node uses the TRA.
After transmitting P, it is monitored whether or not a message indicating that this TRAP has been received normally is received.
If a message of P normal reception is not received, TRAP is repeated and retransmitted until this message is received.

[0008]

However, the above-mentioned prior art has the following problems.

The first problem is that TRAP may drop out during transmission and may not reach the manager, so that all information cannot be reliably reported.

The second problem is that, when polling is periodically performed from the manager, if the interval of the cycle is lengthened, information acquisition on the manager side is not performed in real time. On the other hand, if the interval of the cycle is shortened, the state changes. Without this, the network is constantly loaded, and the network load increases.

The third problem is that T reaches the manager.
If the RAP was not in order, the manager could not recognize that order.

An object of the present invention is to solve the above-mentioned problems,
SNMP T that manager can get TRAP without fail
It is to provide a RAP management method.

[0013]

DISCLOSURE OF THE INVENTION The TR of the SNMP of the present invention
In the AP management method, each time the agent notifies the manager of information on the state change of each node, the version number of the notification to the manager is added to TRAP and notified.
The manager that has received the AP compares the version number added to TRAP with the version number stored by itself before analyzing the information on the node state change due to TRAP, and updates the version number storage if they match. If not, T
RAP retransmission is requested.

The version number added to the TRAP sent from the agent to the manager is two versions: the version number of the previous notification to the manager and the version number of the current notification.

Further, the version number added to TRAP is not limited to the order of serial numbers or chronological order. The version number of the previous notification to the manager and the version number of the current notification to the manager are determined by an arbitrary symbol string or the like. It is only necessary that the two versions be distinguished.

Further, the manager uses the TRAP method and the polling method having a long cycle together, and the manager
Since no RAP is received, detection of an unreceived TRAP may be prevented from being left for a long time.

The SNMP TRAP management device of the present invention comprises:
SNM that handles SNMP protocol to manager
A P processing unit, a main processing unit that performs monitoring control processing,
Previous T received from control program and agent
A memory for storing data including the version number of the RAP; an agent has an SNMP processing unit for processing an SNMP protocol; a main processing unit for monitoring and controlling each node; a control program and a previous manager; It has a memory for storing data including the version number of the transmitted TRAP, and a node communication unit for communicating with each node.

Further, agents and managers may be grouped for each management target item, and one or more communities may exist on the network.

[0019]

Next, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 is a system configuration diagram of one embodiment of the present invention, and FIG. 2 is a sequence diagram of version number management.

In FIG. 1, the system of the present embodiment comprises a node 13 and an agent 12 for monitoring the node 13.
Connected to the agent 12 via the LAN, and
And a manager 11 for remotely monitoring and controlling the agent 12 using the MP.

In the present embodiment, one manager 11
And the agent 12 are connected to the same network on the LAN. However, according to the specification of SNMP, the network between the manager 11 and the agent 12 may be a WAN in which different networks are connected via a repeater or the like. Good. Further, the manager 11 can also perform remote centralized monitoring control of a plurality of agents.

The protocol between the agent 12 and the node 13 can be any protocol.
The agent 12 can monitor and control a plurality of nodes 13.

The manager 11 includes an SNMP processing section 11-1 for processing an SNMP protocol, a main processing section 11-2 for performing monitoring control processing, and a control program and a version number of the previous TRAP received from the agent. An agent 12 has a memory / hard disk 11-3 in which data is stored, and has an SN that processes SNMP.
An MP processing unit 12-1, a main processing unit 12-2 for performing monitoring control processing, a memory / hard disk 12-3 for storing a control program and data including a TRAP version number previously sent to the manager, and a node And a node communication unit 12-4 that communicates with the node 13.

The manager 11 and the agent 12 always store the latest TRAP version information in their respective memories 11-
3, 12-3.

When there is a change in the state of the node 13, the agent 12 notifies the manager 11 of the information by TRAP to which the latest version number and the last version number transmitted by the previous TRAP are added. When receiving the TRAP, the manager 11 compares the version number of the agent 12 secured in the memory 11-3 with the final version number of the TRAP received by the manager 11, and if the version numbers are different, the TRAP is leaked. Is determined, and the manager 11 requests the agent 12 to retransmit a TRAP after the last version of the TRAP received.

Next, the operation of this embodiment will be described.

When a change in state occurs in the node 13, the agent 12 transmits a TRAP regardless of whether the manager 11 is in the active state. Therefore, when the manager is started, before starting normal operation, the latest configuration information and version number information stored in the memory 12-3 of the agent 12 are requested, and these information are acquired. 11-3. At this time, the version number (A1) of the manager 11 and the version number (M
1) is consistent.

While the manager 11 is running and operating normally, when a state change occurs in the node 13, the agent 12 analyzes the state change and notifies the manager 11 of the final version number ( M1) and the latest version number (M2) are notified to the manager 11 as TRAP (S3). Manager 11 receives the TR
The final version number (M1) of the AP (S3) is compared with the final version number (A1) stored in the memory 11-3. In this case, since the two match, the manager 11
The monitoring process is performed by analyzing (S3), and the final version number (A1) of the memory 11-3 is updated to (A2) to have the same value as (M2).

However, if TRAP does not arrive at the manager 11 for some reason, the TRAP (S4) to which the final version number (M2) and the latest version number (M3) are added arrives at the manager 11 from the agent 12, for example. If not, the manager 11 does not update the memory 11-3, so the final version number of the memory 11-3 remains (A2), which is the same value as (M2).

After that, a further state change occurs in the node 13, and the agent 12 sends the TRAP (S5) to which the final version number (M3) and the latest version number (M4) are added to the manager 11, and the manager 11 next issues This TRAP (S
Assume that 5) is received. Manager 11
The final version number (M3) added to (S5) and the memory 11-
By comparing with the final version number (A2) stored in No.3, it is recognized that TRAP (S4) has not been received due to mismatch. Therefore, the manager 11
The retransmission request signal Set-Requ, to which the final version number (A2) stored in the memory 11-3 is added without performing the analysis of (S5).
est (S6) is sent to the agent 12 to request retransmission of TRAP.

After returning a response signal Get-Request (S7) to the retransmission request (S6), the agent 12 retransmits TRAP (S4, S5) after the version number (A2) to the manager 11.

Next, a case where the TRAP arrives at the manager 11 out of order will be described by comparing the conventional system with the system according to the present invention.

First, the manager 11 and the agent 12
Are both in “State 1” and their final version numbers (A1
It is assumed that 1) and (M11) match.

Next, when a state change occurs in the node 13 and the state changes to “state 2”, the agent 12 changes the state to “state 2”.
Is transmitted to the manager 11, and when a further state change occurs in the node 13 and the state changes to “state 3,” the TRAP (S12) indicating that the state has changed from the agent 12 to “state 3”. ) Is Manager 1
Sent to 1. However, if the route that each TRAP passes is different, the order of the TRAPs received by the manager 11 does not always match this transmission order, and may be reversed.

In such a case, according to the conventional method, in such a case, the manager 11 performs processing assuming that the state of the node 13 has changed to “state 1”, “state 3”, and “state 2” in the order of arrival of TRAP.

However, according to the present invention, the manager 11
Analyzes the final version number of the TRAP (S12) of "State 3" received earlier, recognizes the existence of unreceived TRAP, and causes the agent 12 to recognize the TRP after "State 1".
Since the retransmission request of the AP (S11, S12) is made, it is possible to prevent the order of the state change from being reversed.

As described above, according to the present invention, the agent sends the TRAP with the last version number previously sent and the current version number added to the TRAP, and the manager updates the version number stored in the TRAP by the TRAP. Only when the two do not match as compared with the final version number, retransmission of unreceived TRAP is requested, so the manager reports to the agent every time TRAP is received, and the TRAP is performed until the agent confirms the report.
The load on the network is greatly reduced as compared with the conventional example in which the transmission of the AP is repeated.

[0039]

As described above, according to the present invention, two versions of the TRAP version previously transmitted from the agent to the same manager and the version of the TRAP transmitted this time are set to TRA.
By notifying in addition to P, the manager
There is an effect that TRAP can be reliably obtained by detecting leakage of reception of P.

The version number added to the TRAP sent from the agent to the manager is two versions, the version number of the previous notification to the manager and the version number of the current notification. Has the effect of simplifying the processing of.

Also, by setting the number of editions to be added to TRAP by an arbitrary symbol string or the like that can distinguish between the immediately preceding TRAP and the current TRAP for the manager, discontinuous edition numbers are assigned to a plurality of managers. There is an effect that can be.

Further, since the agent uses the TRAP method and the polling method having a long period in combination, the manager does not receive the next TRAP, so that the unreceived TRA is not received.
There is an effect that the detection of P can be prevented from being left for a long time.

[Brief description of the drawings]

FIG. 1 is a system configuration diagram of an embodiment of the present invention.

FIG. 2 is a sequence diagram of version number management in FIG. 1;

FIG. 3 is a normal TRAP sequence diagram.

[Explanation of symbols]

11 Manager 11-1 Main Processing Unit 11-2 SNMP Processing Unit 11-3 Memory / Hard Disk 12 Agent 12-1 Main Processing Unit 12-2 SNMP Processing Unit 12-3 Memory / Hard Disk 12-4 Node Communication Unit 13 Node A1 A13 Manager version M1 to M13 Agent version S1 to S12 Communication between manager and agent including TRAP

Claims (6)

[Claims] A plurality of nodes, an agent for monitoring and controlling the node, and information on a change in the state of the node is notified from the agent using SNMP TRAP which is a network management protocol, and the agent is remotely controlled. A management method for SNMP TRAP transmission in a network management system comprising a controlling manager, wherein the agent transmits information on a state change of the node to T
Each time the manager is notified by RAP, the version number of the notification to the manager is added to TRAP and notified. The manager receiving the TRAP adds the version number to the TRAP before analyzing the information on the state change of the node. The SNMP version is compared with the version number of TRAP stored by itself and updates the version number storage if they match, and requests the agent to retransmit the TRAP if they do not match. TRAP management method. 2. TR sent from an agent to a manager
2. The SNMP TRAP management method according to claim 1, wherein the number of versions added to the AP is two versions: a version number of a previous notification to the manager and a version number of a current notification. 3. TR sent from an agent to a manager
The version number added to the AP distinguishes the two versions, the version number of the previous notification and the version number of the current notification, to the manager by numerical order, chronological order, or any symbol. 3. The SNMP TRAP management method according to claim 1 or 2. 4. The SNMP trap according to claim 1, wherein said manager receives a TRAP to which a version number is added each time a node status change occurs and periodically polls said nodes and agents. Management method. 5. An SNMP processing unit for processing a plurality of nodes, a network management protocol called SNMP, a main processing unit for performing monitoring control processing of each node,
An agent that includes a memory in which a control program and data are stored, and a node communication unit that communicates with each node, generates an SNMP TRAP that notifies information of a change in state of the node, and monitors and controls each node. An SNMP processing unit that processes an SNMP protocol, a main processing unit that performs a monitoring control process, and a memory in which a control program and data are stored. The agent is notified from the agent using SNMP TRAP. In a network management system comprising: a manager performing remote monitoring and control of the agent based on information on a change in the state of a node; wherein the agent generates a version number of TRAP and adds it to TRAP transmitted to the manager; Means for storing the updated version number in its own memory. The manager includes means for storing the version number of TRAP received in its own memory, when receiving the TRAP from the agent, the TR
Means for comparing the version number added to the AP with the version number stored in its own memory to detect a mismatch; and, when the mismatch is detected, the version stored in its own memory for the agent. T sent after number
Means for requesting that the RAP be retransmitted. 6. The SNMP trap management apparatus according to claim 5, wherein the agent and the manager include one or more communities grouped for each management target item.