JP2000201156A - Network management system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a network management system capable of suppressing load to be applied to a network and realized by simple constitution so that the quality of network management is not reduced. SOLUTION: Plural polling weighted lists common to a manager and respective agents are prepared and priority information is stored in each management information to be collected. Prior to poling, the manager informs of a list to be used which is determined by the manager out of plural poling weighted lists. The manager is provided with a poling timer part 73 for counting poling periods and a poling counter part 74 whose counts value is counted up in each poling period and allowed to inform each agent of the count value in each poling period. Each agent collects management information based on the informed count value and priority information stored correspondingly to management information stored in the selected poling weighted list.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a network management system, and more particularly to a network management system for suppressing a network load.

[0002]

2. Description of the Related Art Conventionally, as one of communication protocols of a network such as a computer, a transmission control protocol /
Internet Protocol (Transmission Control Pro
tocol / Internet Protocol: hereinafter abbreviated as TCP / IP. ). The network based on TCP / IP is widely used in universities and industries around the world. As a communication protocol for managing these devices on the network, a simple network management protocol (Simple N
etwork Management Protocol: hereinafter abbreviated as SNMP. ) Has been proposed. The SNMP performs various types of management on management information called a management information base (hereinafter, abbreviated as MIB).
Network management by SNMP is performed by an agent that operates on each managed node in the network and collects and sets management information of the managed node, and a manager that centrally manages management information from the managed nodes. . The manager performs centralized management of management information by transmitting and receiving messages by SNMP with the agents of the managed nodes.

FIG. 12 shows an outline of the configuration of a network management system that performs such network management using SNMP. The network management system includes one management station 11 and a plurality of managed nodes 12 ₁ ,
12 _2,..., It is connected and 12 _N. The management station 11 has a manager (not shown). Each of the managed nodes 12 ₁ , 12 ₂ ,..., 12 _N has an agent (not shown). Manager in the management station 11, the message routing information and transmitting and collecting and setting MIB made of traffic information and failure information such as the number of received messages, each of the managed nodes _{12 1, 12 2, ···,} 12 N (Dashed lines 13 ₁ , 13 ₂ ,..., 13 _{N in} FIG. 12). This request is sent from the manager in the management station 11 by T
Each of the managed nodes 12 ₁ , 12 ₂ ,..., In the form of a request message transmitted over the CP / IP network 10.
12 _Performed for agents in _N. The agents in each of the managed nodes 12 ₁ , 12 ₂ ,..., 12 _N that have received this request collect and set the requested information in each node. Then, the processing result is expressed in the form of a response message to the request message in the form of a TCP / I
The data is sent to the manager in the management station 11 via the P network 10 (broken lines 14 ₁ , 14 ₂ ,.
.., 14 _N ).

[0004] As described above, the network management based on the SNMP is performed according to a request from a manager in the management station 11 by each of the managed nodes 12 ₁ , 12 ₂ ,
..., _12N are performed when the agents in _N return the management information. In addition, each managed node 12 ₁ ,
12 _2,..., Each agent in the 12 _N is to notify the management information in the form of an event to the manager in the autonomous management station 11. This also enables centralized management of management information of the network system. In any case, such information collection and setting by SNMP are performed by an SNMP message having a predetermined format transmitted and received between the manager and the agent. For example, SNM for collecting certain management information from manager to agent
The P message includes an “SNMP GET” message. In addition, S for sequentially collecting management information
"SNMP GETNEX" as an NMP message
T "message or" SNMP GETBUL "as an SNMP message for collectively collecting management information.
A message that is sent from the agent to the manager in response to the request message includes an “SNMP GET RESPONSE” message, etc. Further, the agent informs the manager of management information on an irregular basis. As an SNMP message for generating an event, "SNMP T
RAP "message.

FIG. 13 shows an outline of the format configuration of such an SNMP message. FIG. 2A shows an outline of a format configuration of an SNMP message, and FIG. 2B shows a PDU (Protoc
ol Data Unit) section. As shown in FIG. 2A, the SNMP message format 15 is composed of a version number 16, a community name 17, and a PDU section 18. The version number 16 describes the version of SNMP used as a protocol. Thus, even if the format configuration is changed due to a future version upgrade, it is possible to take a flexible measure such as converting the format configuration or discarding the format configuration as it is. The community name 17 is an identifier for identifying several categories divided as management information. For example, when the version number 16 or the community name 17 does not match between the sending side and the receiving side, the message is discarded as it is. The PDU section 18 describes communication contents according to the message. The PDU section 18 is provided in the same manner as in FIG.
As shown in the figure, PD for identifying the type of message
It is composed of a U type 19, a request_id 20, which is an identifier of a request message from the manager, and a PDU data section 21. By referring to PDU type 19, the message can be referred to as “SNMP G
ET ”message or“ SNMPGETNEX ”
T ”message or“ SNMP GETBUL ”
K ”message or“ SNMP GET RES ”
PONSE ”message or“ SNMP TRA ”
The request_id 20 is an identification number uniquely assigned to identify the request message sent by the manager, and the agent side transmits the same request_id to the response message. The PDU data section 21 differs depending on the message identified by the PDU type 19, and describes, for example, management information 22 or the like.

In the network management using SNMP, the manager in the management station 11
_1, 12 _2, ..., and sends the regular SNMP messages described above for each agent in the 12 _N. This implements polling and collects management information included in the response message.

FIG. 14 shows an example of such an SNMP message, an "SNMP GET" message and a response message "SNMP GET RESPONS".
E "message is transmitted and received. In this case, the manager 23 sends the" SNMP GET "message.
By sending message 24 to agent 25,
A request is made to collect management information managed by the agent 25. Regarding which management information is collected, refer to “SNM
"Specified in the message within 24. That format structure in which the SNMP message shown in FIG. 13" P GET SNMPGET "message 24, the version number 27 _1" 0 ", the community name 27 _2" aab
b ", to request_id27 _3" 23 ", PDU
With specifying the "GET" to type 27 _4, PDU
Describe management information to be collected in the data section. This management information includes a management information object identifier (Object IDent).
ifier: Hereinafter, abbreviated as OID. ), And describes a desired OID. FIG.
In, for example, administration to item 28 _{1-28 3} describes a first to third OID, the management information data unit 29 ₁
29 to ₃ indicate that the description information is meaningless "NUL
L ”is described.

FIG. 15 shows the management information O
It shows an outline of the configuration of an ID registration tree. By managing the MIB as a tree-structured management information group, collection of desired management information is facilitated. In this tree configuration, under “root”, “cc”
The subtree “itt” is registered.
o "subtree," joint-iso-ccit "
The subtree of "t" is registered.
By setting the OID specified from “root” to “0”, the subtree of “ccitt” can be specified.
Similarly, when the OID is set to “1”, the subtree of “iso” is set, and when the OID is set to “2”, “joint-iso-cc” is set.
It means that the subtree of "itt" can be specified, and "std", "reg authority",
The sub-trees of “member body” and “org” are registered, and have a tree structure as a whole. For example, to specify a subtree of "org",
The OID specified from “root” is “1.3”.
Similarly, for example, when "aa" under the subtree of "enterprises" is designated, the OID is "1.
3.6.4.1.4.1. xx ". By adopting such a tree structure, it is possible to easily cope with future expansion.

Returning to FIG. 14, the description will be continued. By describing the first to third OIDs in this manner, an “SNMP GET” message 24 that specifies management information to be collected
Is received by the agent 25. Agent 25
Collects the management information specified in the managed node, and sends a response message “SNMP GET RESPON
SE "responds to the manager 23 as a message. That is," SNMPGET RESPONSE "message 26, the version number 30 _1" SNMP GE
T "message same as" 0 ", the community name 30 _2" SNMPGET "message same as the" aabb ", r
equest_id30 ₃ to "SNMP GET" message same as "23", "GET to the PDU type 30 ₄
RESPONSE ”and describe the collected management information in the PDU data section. The collected management information is stored in the management item data sections 32 _{1 to} 32 ₃ corresponding to the management item sections 31 _{1 to} 31 ₃ , respectively. For example, "123
4 "," XY P5 ", and" AB-UX 5.0 ".
t_id30 reference to ₃ and PDU type 30 _4. This makes it possible to recognize that Request_id27 ₃ is a response to "SNMP GET" message 24 of "23". Like this
By referring to st_id, it is possible to easily recognize which request message is a response message to the request message.

In general, the manager receives the "SNMP TRAP" message sent from the agent at irregular intervals in combination with such polling, and realizes the management information collection function by being notified of the occurrence of an event. It is also done. However, "SNMP
Regarding the TRAP “message”, it is not possible to guarantee that the message has been reliably delivered from the agent to the manager. Therefore, collection of management information is mainly performed by polling.

Various technologies relating to such a network management system have been proposed. For example, JP-A-7-1
No. 83905, entitled "Remote Monitoring System", has an important feature in that a polling level from a management station to a plurality of managed nodes is changed according to a requested data level held in each of the plurality of managed nodes on the network. There is disclosed a technology relating to a network management system that preferentially collects high-level data.

[0012] For example, in Japanese Patent Application Laid-Open No. 8-292922 "Network management device", a manager creates a PDU containing a management attribute value until the maximum PDU size is reached within a range that can be answered by an agent, and uses the PDU. A technology related to a network management system that optimizes management communication by making various requests to agents has been disclosed.

[0013] For example, in JP-A-9-172435, "Distributed Management System", a polling function is provided on the agent side, and the manager is notified only when a state change or abnormality occurs, thereby distributing the network load. A technology related to a network management system has been disclosed.

[0014]

However, in the network management system shown in FIG. 12, the function of the managed node having the agent becomes complicated and the number of management items increases. If the number itself increases, the number of request messages sent from the manager to each agent and the number of response messages from each agent responding to this to the manager increase, and the load on the network increases. Problem.

Further, if the manager sets a longer polling cycle for each agent as a measure against an increase in the network load, the polling interval is determined within the processing capacity of the manager. There is a secondary problem that the manager cannot respond in real time to a managed object whose management items managed in the node change significantly.

Further, in the technique disclosed in JP-A-7-183905, the polling level from the management station to a plurality of managed nodes is changed according to the requested data level held in the managed node. If the transmission level is not changed and notified for each polling, there is a problem that management information below a certain transmission level cannot be collected at all. This lacks reliability in performing network management. In addition, since the load on managers or agents will always be higher than before, with the increase in management items and management targets,
The scalability of the system becomes difficult.

Japanese Patent Application Laid-Open No. Hei 8 (1994) creates a PDU so as to have a maximum PDU size as long as the agent can respond.
In the technique disclosed in Japanese Patent Application Laid-Open No. 29292/1990, there is a problem in that a management object in which a management item managed in a managed node changes significantly has an excessively large load to cope with the change. . Naturally, as the number of items to be managed and the number of objects to be managed increase, a greater load is always applied to the manager or the agent than before, making it difficult to expand the system.

Japanese Patent Application Laid-Open No. Hei 9-172 for providing a polling function on the agent side to distribute the load on the network.
In the technique disclosed in Japanese Patent No. 435, there is a problem that the manager cannot respond in real time to a managed object whose management item managed in the managed node provided with the agent changes significantly. Furthermore, the report from the agent to the manager is "SNMP TR
Since the message is transmitted using the "AP" message, there is no means for confirming that the message has been securely delivered, and there is a problem in reliability.

Therefore, an object of the present invention is to realize a simple configuration so as not to reduce the load on the network by suppressing the number of request messages and response messages by polling from the manager to each agent, and not to degrade the quality of network management. To provide an improved network management system.

[0020]

According to the first aspect of the present invention, (a) storage means for storing management information to be collected and the priority of the collection in association with each other; A requesting collection requesting means, and (c) when the collection requesting means requests the collection of management information, referring to the priority stored in the storage means corresponding to each management information, and requesting by the collection requesting means. The network management system includes management information collecting means for collecting management information determined by the number of times, and (d) management information sending means for sending only the management information collected by the management information collecting means to the collection requesting means. Let it.

That is, in the first aspect of the present invention, the management information to be collected and the priority of the collection are stored in the storage means in association with each other. Then, the collection of the management information is requested by the collection request unit, and the management information collection unit refers to the priority stored in the storage unit corresponding to each management information, and thereby, according to the number of requests by the collection unit. Only the determined management information is collected by the management information collecting means. The management information collected by the management information collecting means is transmitted by the management information transmitting means to the collection requesting means which has requested the collection of the management information.

According to the second aspect of the present invention, (a) clock means for counting a predetermined cycle, (b) counting means for counting up the number of polls for each cycle counted by the clock means, and (c) A collection requesting unit for requesting collection of management information to be collected for each period counted by the clocking unit and transmitting the number of polls counted by the counting unit; and (d) management information to be collected and collection of the management information. Storage means for storing the priorities in association with each other;
(E) When the collection of the management information is requested by the collection request means, for each management information, the remainder value obtained by dividing the priority value indicating the priority stored in the storage means by the number of times of polling is The network management system includes management information collecting means for collecting management information when the number is zero, and (f) management information transmitting means for transmitting only the management information collected by the management information collecting means to the collection requesting means. .

That is, according to the second aspect of the present invention, it is known that the predetermined period is counted by the timing unit, and the number of polls is counted up by the counting unit for each counted period. Then, the collection request unit requests the collection of the management information every period, and transmits the number of polls counted by the counting unit. The storage unit stores in advance the management information to be collected and the priority of the collection in association with each other. Then, the management information collection unit refers to the priority stored in the storage unit corresponding to each management information together with the notification of the number of polls by the collection request unit, and determines a priority value indicating the value by the number of polls. Management information is collected only when it is divisible. The management information collected by the management information collecting means is transmitted by the management information transmitting means to the collection requesting means which has requested the collection of the management information.

According to a third aspect of the present invention, in the network management system according to the second aspect, the collection request unit sets the number of times of polling in a predetermined management object by a predetermined setting request message using a simple network management protocol. Requesting the collection of the management information, and the management information sending means uses a simple network management protocol to the collection request means, and the management information collection means uses a predetermined setting response message corresponding to the setting request message. The collected management information is transmitted.

That is, according to the third aspect of the present invention, the SNM
Using the P, the collection requesting means requests the management information collecting means to collect the management information by a setting request message including the management information to be collected for a predetermined managed object and the number of polls. On the other hand, the management information sending means sends management information collected by a setting response message corresponding to the SNMP setting request message.

According to the fourth aspect of the present invention, (a) the management information to be collected is given an identifier for uniquely identifying the management information and a predetermined setting request message is sent to a predetermined management object using a simple network management protocol. Collection request means for requesting the collection, (b) storage means for storing management information to be collected and the priority of the collection in association with each other, and (c) collection of management information by the collection request means. Management information collecting means for collecting management information determined by the number of requests by the collection request means by referring to the priority stored in the storage means in correspondence with each management information; The management information collected by the collection means is converted into a predetermined setting response message corresponding to the setting request message by using a simple network management protocol. The imparted to and a management information sending means for sending to the collection request unit to the network management system.

That is, in the invention according to claim 4, the management information to be collected and the priority of the collection are stored in the storage means in association with each other. Then, the collection requesting means assigns an identifier uniquely identified to a predetermined setting request message by SNMP to request collection of management information. The management information collection means refers to the priority stored in the storage means corresponding to each management information when requested to collect the management information by the setting request message, thereby determining the number of requests by the collection means. Only the management information to be collected is collected by the management information collection means. The management information collected by the management information collecting means is sent to the SNMP by the management information sending means.
Using the predetermined setting response message corresponding to the setting request message, the same identifier as that of the setting request message is assigned, and the management information is transmitted to the collection requesting unit which has requested the collection.

According to a fifth aspect of the present invention, in the network management system according to any one of the first to fourth aspects, a plurality of storage units are provided, and the collection request unit transmits the management information collection request to the management information collection unit. A storage unit to be selected from the plurality of storage units is specified prior to the request.

That is, according to the fifth aspect of the present invention, a plurality of storage means are provided, and prior to collection of the management information by the collection request means, the storage means to be used is designated from the collection request means. As a result, even if the range of management information to be collected covers a wide range, the storage means can be easily changed as appropriate, and a flexible network management system can be provided.

According to a sixth aspect of the present invention, in the network management system of the first to fifth aspects, the collection requesting means transmits the management information collected after a predetermined time elapses after requesting the collection of the management information. In addition to the storage, the management information that cannot be collected is stored to that effect.

That is, in the invention according to claim 6, the collection requesting means holds the collected management information when all the requested management information is not collected within a predetermined time after the collection of the management information is requested. The management information that is not collected is kept so that it is not collected. As a result, if any error is desired to be included in the management information obtained in response to the request in the conventional SNMP network management system, the SNMP response packet itself is determined to be unreliable, and the All management information was destroyed, but the management information collected correctly can retain its contents,
Efficient data collection can be enabled.

[0032]

BEST MODE FOR CARRYING OUT THE INVENTION

[0033]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described in detail below with reference to embodiments.

FIG. 1 shows an outline of the configuration of a network management system according to an embodiment of the present invention.
In this network management system, one management station 41 and a plurality of managed nodes 42 ₁ , 42 ₂ ,..., 42 _N are connected via a network 40. In the network 40, communication by TCP / IP is performed. The management station 41 includes a manager (not shown). Each managed node 42 ₁ , 42 ₂ ,
.., 42 _N each have an agent (not shown). The manager in the management station 41
The collection and setting of the MIB including the message route information, the traffic information such as the number of transmitted and received messages, and the failure information are performed by the managed nodes 42 ₁ , 42 ₂ ,.
Ask 42 _N This request is sent from the manager in the management station 41 to each managed node 4 in the form of a request message sent over the network 40.
2 _1, 42 _2,..., Made to the agents in the 42 _N. Each of the managed nodes 42 ₁ , 42 ₂ ,
.., 42 _N perform the collection and setting processing of the requested information in each node, and report the processing result in the management station 41 via the network 40 in the form of a response message to the request message. To the manager.

The network management system according to this embodiment includes a management station 41 and each managed node 42.
_1, 42 _2, ..., a plurality of polling weighted list that is weighted based managed objects managed by an agent, respectively in 42 _N on the frequency and severity of change corresponding to the situation of that value. At the time of polling, one of the plurality of polling weight lists is selected, and each object is weighted with reference to this list. Then, in a simple manner, the polling cycle for the management information with high priority is shortened,
It is characterized in that the number of transmitted / received messages is reduced by making the polling cycle longer for management information with low priority.

Further, each agent of the network management system in the present embodiment manages the selection of one of the plurality of weighted lists when the MIB is selected.
An object and an MIB object that triggers polling are provided. With respect to these MIB objects, an appropriate value can be set by a manager using an “SNMP SET” message. For the MIB object that manages the selection of the list, a desired polling weight list can be selected by setting an appropriate value in the “SNMP SET” message specifying the list to be selected prior to polling from the manager. . The agent returns a "SNMP GET RESPONSE" message in response. For the MIB object that triggers polling, the manager sets an appropriate value in the "SNMP SET" message in order to collect management information in the agent, so that a polling request can be made as a result. It has become. In response to the MIB object set to an appropriate value in the “SNMP SET” message, the agent sends the value of the management information collected in the “SNMP GETRESPONSE” message to the manager as a response. In this case, the "SNMP SET" message for the MIB object that triggers polling returns a "SNMP GET RESPONSE" message as a response. As described above, by transmitting and receiving the general “SNMP SET” message and “SNMP GET RESPONSE” message between the manager and the agent, the selected polling weight list is selected, and the polling considering the priority is performed. I have.

FIG. 2 shows such an “SNMP SET” message and its response message “SNMP GE”.
This shows an example of transmission and reception of a “T RESPONSE” message.
By sending an “MP SET” message 46 to the agent 47, it is possible to set a desired value in a managed object managed by the agent 47. For determining which value is set in which managed object, the “S
"Specify within message 46. That SNMP messages in a format configuration of that shown in FIG. 13" NMP SET SNMP SET "message 46, the version number 48 _1" 0 ", the community name 48 _2" x
xyy 4 "", in request_id48 ₃ ", PD
With specifying the "SET" to the U-type 48 _4, PD
The identifier of the management object to be set and its set value are described in the U data section. This management object has a plurality of management items identified by the management OID, and describes a desired OID managed by the agent. In Figure 2, for example control items 49 _1-49 The ₂ first and second
OID of the management information data section 50 ₁
50 The ₂ describes the "stu" and "15.21.49" is a set value of the first and second OID respectively.

The “SNMPSET” message 46 specifying the set values for the first and second OIDs as described above is received by the agent 47. The agent 47 assigns “stu” and “1” to the management object identified by the first and second OIDs, respectively.
5.21.49 ”, and returns a“ SNMP GET RESPONSE ”message 51 as a response message to the manager 45. That is,
“SNMP GET RESPONSE” message 5
1, the same as the "SNMP SET" message to the version number 52 ₁ "0", the community name 52 ₂ "SN
“Xxyy”, req same as “MP SET” message
uest_id52 ₃ to "SNMP SET" the same as the message "4", the PDU type 52 ₄ "GET RE
As well as specify sponse ", managed object set in the PDU data part and describing the settings. Managed objects this setting, management items 53 _1, 5
3 to the management item data section 54 _1, 54 ₂ corresponding to _2, each "stu", "15.21.49" is described.
The manager 45 that has received this request_request_i
By referring to the d52 ₃ and PDU type 52 _4, request_id48 ₃ is "4""SNMP
It can be recognized that this is a response to the “SET” message 46. In this way, the request_id
, It is possible to easily recognize which request message is a response message.

FIG. 3 shows how an SNMP message is transmitted and received between a manager and an agent in the network management system according to the present embodiment using such an SNMP message. The manager in FIG. 3 is the management station 41 in FIG.
First, the agent is the managed node 42 ₁ in FIG.
42 _2, ..., to 42 _N, respectively correspond. Before starting the polling, the manager determines a polling weight list to be used from among a plurality of polling weight lists that the agent has in advance (process 55 in FIG. 3). Then, in order to notify this to the agent side, an appropriate value is set in the “SNMP SET” message 56 for the MIB object that manages the selection of the polling weight list in the agent. This “SNMP SET” message is transmitted as shown in FIG.
, And data that can be identified as a “SNMP SET” message is described in the PDU type. The agent selects one of a plurality of polling weight lists based on the set value (process 57 in FIG. 3). Thereafter, the agent sends a “SNMPGET RESPONSE” message to the manager as a response (response 58 in FIG. 3).

Subsequently, an appropriate value is set by the "SNMP SET" message to the MIB object associated with the trigger setting so as to trigger polling from the manager (setting 59 in FIG. 3). Upon receiving this, the agent sets a trigger by setting its value for the corresponding MIB object (FIG. 3).
Processing 60), the management information is set in the SNMP level packet as a response to the “SNMP GET RESP
The message is sent to the manager as an “ONSE” message.
ONSE "message, its request_id
We make the following ingenuity. That is, the agent refers to the contents of the polling weight list to be used, and reques the value set by the manager as it is.
est_id field (process 61 in FIG. 3),
"SNMP GET" which is an SNMP level packet
RESPONSE "message 62 to the manager. Even if there are a plurality of packets at the SNMP level, the same request is applied to all the packets.
The value of _id is used. And SNMP PDU re
By extending the concept of the quest_id field,
The manager can receive a polling response message consisting of many packets at the SNMP level by triggering one packet of the "SNMP SET" message. As described above, the manager sets a predetermined value in the “SNMP Set” message for the MIB object serving as one polling trigger in the agent, so that the management information, which is a value held by the MIB object one after another, is changed. Will be replied to.

The manager receiving the reply from the agent confirms whether the response as requested is obtained by comparing the contents indicated by the polling weight list with the contents of the received packet. Step 63 of 3). Then the manager
The contents of the management information received by the polling are displayed on the screen (process 64 in FIG. 3).

Hereinafter, the main components of the network management system according to the present embodiment will be described with reference to FIGS.

FIG. 4 shows an outline of a main part of the configuration of the management station 41 in this embodiment. The management station 41 is a manager application program (Applic) that manages collection of management information by polling for network management in the management station 41.
ation Program: hereinafter abbreviated as AP. ) 70. In addition, an SN that performs an interface with the SNMP
MP—Application Programming Interface (hereinafter A)
Abbreviated as PI. ) 71, and a polling API extension unit 72 for processing a request_id field of the SNMP level to be replied to. Further, a polling timer unit 73 that counts a polling cycle according to an instruction from the manager AP 70, a polling counter unit 74 that holds the number of times of polling each agent, and a plurality of polling weight lists 75 are provided.

FIG. 5 shows an outline of the configuration of the polling weight list in this embodiment. Here, the configuration of one polling weight list is shown. That is, the polling weight list 7
Reference numeral 6 denotes a list in which management information 77 as MIB managed by the agent and its priority information 78 are stored in association with each other. The priority information 78 is information determined in consideration of the frequency of change of the management information 77 and the importance of collecting the management information 77 itself. Here, a smaller number indicates a higher priority.

FIG. 6 shows a manager AP 70 for controlling each functional unit in the management station 11 to perform polling.
3 shows an outline of the processing contents. First, the manager AP 70 performs a plurality of polling weight lists 75 which are held in common with the agent prior to polling.
It is determined which polling weight list is to be referred to from the list (step S80). This is because polling weights created respectively include the time when management information is collected by the manager AP 70, the configuration status of the network system when there is a change in the network to be managed, and the maintenance that periodically performs management. From the list, an alternative is selected according to the situation. When the list to be used is determined in step S80, “SNMP SET” including a predetermined value is set.
The agent is notified using a message (step S81). Next, in order to perform polling at a predetermined cycle, the polling timer unit 73 starts counting time at this cycle and waits for a timeout (step S82:
N). Eventually, when the timeout indicating the polling cycle occurs (step S82: Y), the value of the counter corresponding to the agent to be polled in the polling counter unit 74 is counted up by "+1" (step S83). And this polling A
The count value counted up in step S83 is set to the PI extension unit 72 (step S84). As a result, the polling API extension unit 72 sends the SNMP-A
The request_id field of the SNMP level packet received via the PI 71 is set so as to receive only the same packet as the set count value.

Next, the manager AP 70 sets the count value counted up in step S83 to the agent to start polling (step S8).
5). This is the “SN” including the count value as described above.
The agent receives this and refers to the polling weight list notified in step S81, and replies the management information to the manager based on the count value. This reply is described above. Like "SNMP GET RESP
ONSE "message and its requests
The t_id field is in the form of a packet of the SNMP level set to the count value notified in step S85. Manager AP70 is SNMP-A
Such SNMP-level packets are sequentially received via the PI 71, but the "SNMP SE
When a predetermined time has elapsed after the transmission of the "T" message, it is determined whether or not the management information replied by the agent matches the management information indicated by the selected polling weight list without excess or deficiency (step S
86). Here, when it is determined that all the management information to be collected within a certain time has been replied (step S86:
Y) holds all the management information (step S8).
7). On the other hand, when it is determined in step S86 that all the management information to be collected within a predetermined time has not been replied (step S86: N), the acquired management information is held (step S88), while the acquired management information is retained. The failed management information is stored together with the error information (step S89). As this error information, the fact that the reply was not made within a certain time is held. Note that the management information that could not be obtained in step S89 was replied from the agent, but the content of the management information to be collected due to a failure on the agent side was not received. You may hold | maintain as error information with the content of the error. Thus, steps S87 and S89
After holding the management information and the like, the process returns to step S82 again and waits for a predetermined polling cycle.

When it is desired to change the management information to be collected by polling during such a process, a similar notification can be made by sending a notification of selection of the polling weight list again using the "SNMP SET" message as in step S80. Management information can be collected.

Next, the SNM from such a manager
An agent that receives the P message will be described. The managed nodes 42 ₁ , 42 ₂ ,... In FIG.
42 _N because each identical configuration, will be described below the managed node 42 _1.

FIG. 7 shows the managed node 42 in this embodiment.
₁ shows an outline of the configuration of FIG. Managed node 4
2 _1, the agent 90 ₁ for collecting management information in the management node 42 ₁ at the request by the polling from the manager et al for network management, SNM
SNMP-API 91 for interfacing with P
_{With one} . Further, a plurality of polling weighted list 92 _1, manager A management station 11
By "SNMP SET" message from P70 polling weighting list selector 9 for selecting one list alternatively from these polling weighted list 92 ₁
3 _1, and a polling trigger portion 94 ₁ trigger polling is set by "SNMP SET" message from the manager AP70 management station 11. Polling weighted list 92 ₁ is the same as that shown in FIG.

[0050] Figure 8 shows the outline of the processing contents of the agent for collecting management information of the managed node 42 ₁ in response to a request from the manager AP70 management station 41 controls each functional unit within the managed node 42 ₁ It is a thing.
First Agent 90 _1, in order to select one of a plurality of polling weighted list 92 _1, the step S81 management is requested from the manager AP70 station 41 "SNMP SET" message SNMP-API91 ₁ in FIG. 6 (Step S95: N). And this "SNMP
"When a message is received via the SNMP-API91 _1: (Step S95 Y), the received" SET on the basis of the setting value included in the SNMP SET "message, multiple polling weighting list selection section 93 ₁ alternatively to select one polling weighted list from the polling weighted list 92 _1. Then, taking the contents of the list which is selected in the resource of the agent (step S96). Then, step S85 of FIG. 6 A of the management station 41 at
"SNMP SET" message is requested from the P70 is waiting to be received via the SNMP-API91 ₁ (step S97: N). When receiving this (step S97: Y), received from the management station 41 to the polling trigger unit 94 ₁ "SNMP SE
The count value included in the "T" message is set (step S98).

The count value set here is the number of polls that indicates the current poll request is for many times from the polling start, the agent 90 ₁ this as a trigger for carrying out the collection of management information Become.
So, for each management information to be collected, the agent
Dividing the polling count set in the polling trigger unit 94 _1, in the priority information value as the priority information stored in correspondence with the management information to be collected with reference to polling weighting selected list is notified in step S95 (Step S99). Then, it is determined whether or not the remainder is “0” (step S100). When the remainder is "0": (step S100 Y), the management information, it is determined that the time to be collected, the agent 90 ₁ accesses the system resources corresponding to the management information,
The management information is acquired (Step S101). In this way, for each management information to be collected, the management information to be collected is collected only when the number of polling requests determined by the priority information value matches the number of polls set by the manager.

Then, a reply message is created to reply the management information to the manager (step S102). This reply message is packed in an SNMP-level packet as the above-mentioned "SNMP GET RESPONSE" message, and the request_id field sets the number of polls set in step S98 as it is. "SNMP GET RESPONSE" created in this way
The message is sent to the management station 41 which is the manager.
(Step S103). Even when there are a plurality of reply packets, the same polling count is set in the request_id field of all the packets.

Thereafter, it waits for the polling trigger to be performed again by the manager. Step S100
When the remainder is not "0" (step S100: N)
However, it is determined that the management information is not collected by the trigger set this time, and waits again for the polling trigger from the manager.

Next, the above-mentioned polling priority information will be specifically described.

FIG. 9 shows an outline of a main part of the configuration of the network management system according to the present embodiment described above. The same reference numerals are given to the same components in the respective components described so far. The network management system according to the present embodiment, via the network 40,
One management station 41 and a plurality of managed nodes 4
2 _1, 42 _2, ..., and 42 _N are connected. In the network 40, communication by TCP / IP is performed. The management station 41 includes a manager AP 70. Each managed node 42 ₁ , 42 ₂ ,..., 4
Each of 2 _N has agents 90 ₁ , 90 ₂ ,.
90 _N. The management station 41 and the managed nodes 42 _{1 to} 42 _N are the same as those in FIGS.

FIG. 10 shows an example of one of a plurality of polling weight lists provided in the management station 41 and the managed nodes 42 _{1 to} 42 _N. The polling weight list 104 is a list in which management information 105 as MIB managed by the agent and its priority information 106 are stored in association with each other. The priority information 106 is information determined in consideration of the frequency of change of the management information 105 and the importance of collecting the management information 105 itself. Here, a smaller number indicates a higher priority. Therefore, the priority information of the management information A is “1” and therefore has the highest priority, and the priority information of the management information D and E is “5” and thus has the lowest priority. Here, “high priority” means that management information is collected most frequently in polling, and “low priority” means that the management information collection interval is long.

Here, a description will be given on the assumption that the polling weight list shown in FIG. 10 is selected by the manager and selected by the agent.

FIG. 11 shows how polling is performed for each management information in the network management system according to the present embodiment. That is, it indicates a polling interval that is performed over time for each of the management information A to F illustrated in FIG. As described with reference to FIG. 6, the count value counted up by the polling counter unit 74 is incremented by the polling counter unit 74 for each time T of the polling cycle 107 measured by the polling timer unit 73 in step S82, and an “SNMP SET” message is sent to the agent in step S85. Will be notified. The notification of the count value means a trigger of polling the agent. Therefore, as described with reference to FIG. 8, the agent determines whether to collect management information from the notified count value and the priority information stored in the selected polling weight list.

To explain with the elapse of time, first, polling is started between the manager and the agent, and when the first polling period T is reached, the count value “1” is notified. The agent divides this count value “1” by the priority information as described in step S99 of FIG. In this case, since only the management information A is "0", only the management information A is collected as shown in step S100 of FIG. Therefore, when the polling period T has elapsed, only the collected management information A is “SNMPG”.
ET RESPONSE "message, request_id is described with a count value of" 1 ", and is replied to the manager.

When the polling period 2T has elapsed, a count value "2" is notified. Therefore, the count value “2”
Is divided by the priority information corresponding to each management information, only the management information A and B become "0". That is, when the polling cycle 2T has elapsed, only the collected management information A and B are captured in the “SNMP GET RESPONSE” message, and request_id is described to the count value “2” and is replied to the manager. As described above, the management information is collected at the polling interval based on the count value which is the number of times of polling for each management information. For example, when the priority information is the management information A of "1", at every polling cycle T, when the priority information is the management information B of "2", at every polling cycle 2T, and when the management information C is the priority information of "3" Every 3T polling period,
When the priority information is the management information D of “5”, the polling cycle is 5
For each T, collection is performed at every polling cycle 4T when the priority information is the management information E of “4”, and at every polling cycle 5T when the priority information is the management information F of “5”.

As described above, the network management system according to the present embodiment is provided with a plurality of polling weight lists common to the manager and each agent. In this polling weight list, priority information is stored for each collected management information. Prior to polling, the manager first informs the MIB object that performs the list selection of an SNMP object in order to notify the list to be used determined by the manager from the plurality of polling weight lists before polling. A "SET" message is sent. The agent receives this and selects a use list. Next, the manager includes a polling timer unit 73 that counts a polling cycle and a polling counter unit 74 that counts up every polling cycle.
The count value counted up in step 4 is referred to as “SNMP S
ET ”message.
The count value notified by the “SNMP SET” message is divided by the priority information value stored corresponding to each management information stored in the selected polling weight list. Only when the remainder is “0”, it is determined that the management information is to be collected, the corresponding resource is accessed to collect the management information, and the reply is sent to the “SNMP”.
GET RESPONSE ”message. At this time, the notified count value of the request_id field of the message is used.If the remainder is not“ 0 ”, the management information is not collected. With this configuration, a polling request from the manager to the agent can be realized by one packet at the SNMP level, so that traffic for network management can be reduced. Since the response to the manager can be made based on the priority information of the polling weight list, the number of responses can be greatly reduced, and the traffic can be further reduced.
For high-priority management information, the polling cycle can be kept short, so that it is possible to realize network management that can respond in real time without reducing traffic and without deteriorating the quality of network management. .

In the network management system of this embodiment, management information is collected for a plurality of managed nodes. However, there is a case where a management station and a managed node have a one-to-one correspondence. Is also good.

The configuration of the polling weight list and the items and the number of the list in the network management system according to the present embodiment are not limited to the items.

[0064]

As described above, according to the first aspect of the present invention, the priority is determined by referring to the priority stored in the storage means corresponding to the management information requested to be collected. Only the management information requested by the number of times of request by the collection requesting means is transmitted by the management information transmitting means. This allows the management information collection means to collect the management information based on the priority stored in the storage means when the collection is requested, so that it is not necessary to specify the management information from the collection request means, and polling is performed. In this case, the traffic volume can be reduced. In addition, the management information sending means does not need to respond each time a request is made from the collection request means, but only responds to several requests once according to the priority stored in the storage means. Therefore, traffic between the collection requesting means and the management information sending means can be significantly reduced. Further, the priority stored in the storage unit in association with the management information is set to a high priority for the management information that changes rapidly, and a low priority for the management information that changes little, so that the traffic is reduced. Necessary management information can be collected while suppressing the amount, and efficient network management can be performed.

According to the second aspect of the present invention, the management information collection unit refers to the priority stored in the storage unit in correspondence with the management information requested to be collected together with the notification of the number of polls by the collection request unit. Then, management information is collected only when the priority value indicating the value is divisible by the number of polls. As a result, the traffic between the collection request unit and the management information sending unit can be reduced with a very simple configuration, and the load on the management information collection unit can be reduced.

Further, according to the third aspect of the present invention, the SN
The MP can be used, and can be applied to the network management of a TCP / IP network that is widely spread around the world as one of the communication protocols of a network such as a computer.

Further, according to the fourth aspect of the present invention,
Even when the data amount of the management information sent from the management information collection unit increases and the data is divided into a plurality of setting response messages and collected, the collection request unit that has made the collection request is collected from a plurality of collection destinations. The management information can be easily determined. For example, collection of a plurality of packets at the SNMP level can be realized with a very simple configuration, thereby facilitating quick grasp of management information.

Further, according to the fifth aspect of the present invention, even if the range of management information to be collected is diversified, the storage means can be easily changed appropriately and a flexible network management system can be provided. Can be provided.

Further, according to the invention of claim 6, the collection requesting means holds the collected management information when all the requested management information is not collected within a fixed time after the request for collection of the management information is made. In the case of management information that is not collected, the fact that the information is not collected is retained. As a result, if any error is desired to be included in the management information obtained in response to the request in the conventional SNMP network management system, the SNMP response packet itself is determined to be unreliable, and the Although all the management information has been discarded, the contents of the management information collected correctly can be retained, and efficient data collection can be performed.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating an outline of a configuration of a network management system according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing how a “SNMP SET” message is transmitted and received.

FIG. 3 is an explanatory diagram showing how an SNMP message is transmitted and received in the embodiment.

FIG. 4 is a block diagram illustrating an outline of a main configuration of a management station according to the embodiment.

FIG. 5 is an explanatory diagram illustrating an outline of a configuration of a polling weight list according to the present embodiment.

FIG. 6 is a flowchart showing an outline of processing contents of a manager AP in the embodiment.

FIG. 7 is a block diagram illustrating an outline of a configuration of a managed node in the embodiment.

FIG. 8 is a flowchart showing an outline of processing contents of an agent in the embodiment.

FIG. 9 is a block diagram illustrating an outline of a main part of a configuration of the entire network management system in the embodiment.

FIG. 10 is an explanatory diagram illustrating a specific example of a polling weight list according to the present embodiment.

FIG. 11 is an explanatory diagram showing how polling is performed in the embodiment.

FIG. 12 is an explanatory diagram illustrating a configuration of a conventionally proposed network management system using SNMP.

FIG. 13A is a diagram illustrating a format configuration of an SNMP message. FIG. 4B is a format configuration diagram of a PDU unit.

FIG. 14 is an explanatory diagram showing transmission and reception of an “SNMP GET” message.

FIG. 15 is an explanatory diagram showing a configuration of a registration tree of management information OID.

[Explanation of symbols]

40 network 41 management station 42 ₁ , 42 ₂ ,..., 42 _N managed node 70 manager AP 71, 91 ₁ SNMP-API 72 polling API extension unit 73 polling timer unit 74 polling counter unit 75, 92 ₁ polling weight list 90 ₁ agent 93 ₁ weight list selection unit 94 ₁ polling trigger unit

 ──────────────────────────────────────────────────続 き Continued on the front page F-term (reference)

Claims (6)

[Claims] A storage unit for storing management information to be collected and a priority of the collection in association with each other; a collection request unit for requesting collection of management information; and a request for collection of management information by the collection request unit. Management information collection means for collecting management information determined by the number of requests by the collection request means by referring to the priority stored in the storage means in response to each management information; A management information sending unit for sending only the management information collected by the management unit to the collection requesting unit. 2. Time counting means for counting a predetermined cycle, counting means for counting up the number of polls for each cycle counted by the time counting means, management information to be collected for each cycle counted by the time counting means Collection requesting means for requesting collection and sending out the number of polls counted by the counting means; storage means for storing management information to be collected and the priority of the collection in association with each other; When the collection of the management information is requested, the management information is collected when the remainder value obtained by dividing the priority value indicating the priority stored in the storage means by the number of times of polling corresponding to each management information is zero. Management information collecting means, and management information sending means for sending only management information collected by the management information collecting means to the collection request means. Network management system characterized by Bei. 3. The collection request unit requests the collection of the management information by setting the number of times of polling in a predetermined management object by a predetermined setting request message using a simple network management protocol, and sends the management information sending unit. Transmitting the management information collected by the management information collecting means to the collection requesting means by a predetermined setting response message corresponding to the setting request message using a simple network management protocol. The described network management system. 4. A collection request means for assigning an identifier for uniquely identifying management information to be collected and requesting the collection of a predetermined management object by a predetermined setting request message using a simple network management protocol. A storage unit for storing management information to be collected and the priority of the collection in association with each other; and when the collection request unit requests collection of the management information, the storage unit stores the management information in correspondence with each management information. Management information collection means for collecting management information determined by the number of requests by the collection request means with reference to the priority, and setting the management information collected by the management information collection means using a simple network management protocol Management information to which the identifier is added to a predetermined setting response message corresponding to the request message and which is sent to the collection requesting means A network management system comprising transmission means. 5. A storage means comprising a plurality of storage means, wherein the collection request means designates a storage means to be selected from the plurality of storage means prior to a request for collection of the management information to the management information collection means. Claim 1 characterized by the following
The network management system according to claim 4. 6. The collection requesting means holds the management information collected after a predetermined time has passed after requesting the collection of the management information, and holds the fact about the management information that cannot be collected. The network management system according to claim 1, wherein: