JP3292349B2 - How to grant management rights for a decentralized management agreement type platform - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a connection between a database group and an application program group and a connection between various systems in the management and operation of a network used in a communication field, an office field, and the like. The present invention relates to a method for granting the management right of a distributed management function agreement type platform.

[0002]

2. Description of the Related Art A conventional technique will be described below with reference to FIGS. FIG. 2 shows an application program AP.
₁ and AP ₂ perform a series of tasks using databases DB _A and DB _B storing data groups 1 and 2, respectively. As shown in FIG. 1, in a conventional configuration, a database (hereinafter, referred to as DB) is used.
And an application program (hereinafter, referred to as an AP) that performs a series of tasks using the corresponding DB.

For this reason, when it is difficult to distribute the data of the DB or when the DB is changed, it is necessary to change the AP. That is, the AP ₁ cannot perform a series of tasks using both the data group 1 of DB _A and the data group 2 of DB _B , and it is necessary to create a new AP or system.

[0006] As shown in FIG. 3, in a task module M that performs a specific task, the AP and the task module M are directly connected. Therefore, the task module M
Was changed, the AP also had to be changed. Further, the change in the case of an AP using a plurality of task modules and a plurality of DBs is very complicated.

As a means for solving such a problem, as shown in FIG. 4, an AP and a DB, and an AP and a task module M are separated without being directly connected, and a platform (hereinafter, referred to as a PF and a PF) is a connecting portion for connecting these. ), It is possible to distribute data and reduce the influence of changes in the DB and the task module M on the AP.

The PF comprises an AP interface 1 for interfacing a plurality of APs _{1 to 4} , an M interface ₂ for interfacing a plurality of task modules M ₁ and M _{2, and} an interface for a plurality of DBs ₁ and ₂ . And an accessor 4 that manages access requests from a plurality of AP groups, task modules M, and DBs via the respective interfacers.

Here, the PF is software for connecting a plurality of DBs and task modules M to an AP. For example, PF is a journal of the Institute of Electronics, Information and Communication Engineers, vol.E76-B, No.4, April 1993,
pp.391-401 N.Kashima et.al. "Optical cable network
operation in subscriber loops ".

However, as shown in FIG.
In a situation where B, the task module M, and the AP are connected via the PF, there is a risk that the AP may concentrate on the PF.
As a result, the waiting time for work becomes longer,
If the load is not distributed among the PFs, there is a possibility that the PF may fail due to an excessive load in the worst case, resulting in an inefficient system configuration.

[0009] The PF failure has a large effect on the entire system. That is, the DB and the task modules M and A
Although the failure of the P is local, the failure of the PF affects all DBs, task modules M, and APs connected to the PF, and may have a configuration that affects the reliability of the system.

In general, an N-fold redundant system in which a plurality of identical systems are connected in parallel is employed as a means for improving reliability. However, even if some kind of load distribution means is not applied here, the load concentrates only on a specific PF, which lowers the processing speed of the entire system, induces a failure due to excessive load, and causes a light load state (free state). Including PF).

For this load management control, software having a central management control means called a system switcher as shown in FIG.
Platform Reliable Configuration "IEICE Spring Conference B-
990 (1994).

The system switcher is composed of individual units P
The load status of each unit PF is recognized based on the load information from F, and the load is distributed so that the processing efficiency is maximized in the entire system. Also, the unit PF breaks down,
In order to avoid a delay in the processing of the job requested by the unit PF, a unit PF with a light load is selected from other units PF on the network to continue the job.

[0013]

However, when considering the reliability of the entire system, when a centralized control function such as a system switcher fails, a load is applied only to a specific unit PF in the long term. Concentrate and affect the entire system. Further, the continuation of the backlogging operation which the failed unit PF has been processing is not efficiently executed, so that the waiting time of the operator may be extended.

SUMMARY OF THE INVENTION In view of the above problems, an object of the present invention is to newly add a unit PF in a lightly loaded state when a unit platform to which the management right has been assigned falls into an overloaded state or when a trouble occurs due to a failure. It is an object of the present invention to provide a management right grant method for a decentralized management function agreement type platform for granting a management right to a company.

[0015]

The present invention SUMMARY OF] is in claim 1 in order to achieve the above object, a database group, using task modules for performing specific tasks, and said group of databases and task modules series Management program in a distributed processing system composed of a plurality of unit platforms which are connection units for connecting the database group, the task module group, and the application program group via a network. A method for granting management rights of a type platform, wherein communication for load distribution and operation status diagnosis is periodically performed between a unit platform to which management rights have been granted and another unit platform on a network, Units for which administrative rights have been granted When the communication is interrupted from the platform to another unit platform on the network even after a predetermined time, the unit platform to which the management right has not been assigned becomes the unit platform to which the management right has been assigned. A command for performing its own failure diagnosis is transmitted to the platform, and thereafter, when there is no response from the unit platform to which the management right is assigned, each unit platform to which the management right is not assigned is assigned to its own platform. Diagnose the operating status and load status such as the priority of the application program to be processed, the task module in use, and the average value of the operation time, and notify the information to other unit platforms on the network. Each unit platform that received the notification is notified. Method of selecting a unit platform in a lightly loaded state from information received, determining a unit platform to which a new management right can be assigned, and assigning a management right to the unit platform. Suggest.

[0016]

According to the first aspect of the present invention, communication for load distribution and operation status diagnosis is periodically performed between the unit platform to which the management right is assigned and another unit platform on the network. When the communication is interrupted after a preset time from the unit platform to which the management right is granted to another unit platform on the network, the unit to which the management right is not granted The platform sends an instruction for performing a self-diagnosis to the unit platform to which the management right has been given. Thereafter, when there is no reply from the unit platform to which the management right has been granted to the unit platform that has transmitted the command, each unit platform to which the management right has not been granted has an application program to be processed by itself. The operating status and the load status such as the priority, the task module in use, and the average value of the operation time are diagnosed, and the information is notified to other unit platforms on the network. As a result, the unit platform in the light load state is selected from the notified information by each unit platform that has received the notification, the unit platform to which the management right can be newly assigned is determined, and the management right is assigned to the unit platform. Is performed.

[0017]

An embodiment of the present invention will be described below with reference to the drawings. FIGS. 1 and 6 show the unit P in the present embodiment.
It is a block diagram showing F. The unit PF of the present embodiment has a configuration of a conventional unit PF, that is, in addition to a processing unit 20 including an AP, M, DB interface function 21, an access function 22, and another system interface function 23,
A management control unit 30 having a management / control function is provided.

As shown in FIG. 6, the processing unit 20 includes an AP interface 1 for interfacing with a plurality of AP ₁ and AP ₂ , an M interface ₂ for interfacing with a plurality of modules M ₁ and M 2, DB
₁ , a DB interface 3 for interfacing with the DB ₂ , and a plurality of APs through each of the interfacers
An accessor 4 for managing access requests from the group, the task module M, and the DB is provided.

As shown in FIG. 6, each unit PF has a management control unit 30 including a connection switching function unit and a selection function unit, which are control functions conventionally held by the central management control function.

That is, the management control unit 30 is composed of a unit PF management function unit 31 and a unit PF control function unit 32. The unit PF management function unit 31 diagnoses a failure of the unit PF itself. A failure reporting function unit 6 for reporting a failure of the unit PF itself, and a unit PF
An operation status diagnosis function unit 7 for checking the processing status in the unit, an operation status notification function unit 8 for notifying the processing status in the unit PF, and the failure information from the failure notification function unit 6 and the operation status notification function unit 8 The operation status information is supplied to an operation status information storage unit 9 in the system.

The unit PF control function unit 32 in each unit PF includes a selection function unit 10 and a connection switching function unit 1.
1. The other unit PF failure diagnosis command transmission function unit 12 is provided. These control the load distribution of all the units PF in the network.

Further, each unit PF, AP and task module M are connected on a data transfer network (DCN) such as a LAN as shown in FIG.

According to the distributed management function agreement type platform having the above-described configuration, each unit PF is connected to the network (DCN) as shown in FIG. Diagnosis of the situation is performed, and the information is notified to the unit PF to which the management right is given. As a result, the unit PF to which the management right has been assigned performs load distribution and risk distribution based on the information.

The unit P to which the management right is given
F, as shown in FIG. 9, operates the failure diagnosis function unit 5 periodically to diagnose its own failure state (SA1, SA
2). As a result, if it is determined that a minor failure has occurred, the unit PF to which the
Broadcasts information (relief information) for relinquishing the management right to all the units PF on the network by the failure report function, selects the unit PF that is optimal to newly grant the management right (SA3), The management right is transferred to the unit PF to which the management right is assigned (SA4).

Further, the unit PF to which the management right is given operates the operation status diagnosis function unit 7 periodically to diagnose the operation status (SA5), and the diagnosis result exceeds the preset load standard. In the case of an overload state, for example, when the load value obtained as a result of the diagnosis exceeds 0.5, the process of SA3 is performed (SA6).

That is, the unit P to which the management right is given
F broadcasts information for relinquishing the management (relief information) to all units PF on the network by the operation status reporting function unit 8. Each unit PF receiving this diagnoses its own failure status and load status, and when it determines that its own load status is lighter than the predetermined load standard during normal operation, the load information Is transmitted to the unit PF to which the management right is given. Conversely, if the result of the diagnosis indicates that the load is higher than the reference, only normal operation is continued.

The unit PF in the overloaded state and given the management right compares the received load information of each unit PF, selects the unit PF in the lightest load state, and gives the unit PF the management right. After the transfer, only the operation of the normal unit PF is performed.

Further, the unit PF to which the management right has been granted notifies the last transfer of the management right to all the units PF on the network, or the unit PF to which the management right has been newly provided receives the management right. The right is newly granted to all the units PF on the network.

On the other hand, if the unit PF to which the management right has been assigned cannot perform any operation due to a serious failure, the unit PF to which the management right has been assigned is transferred from the unit PF to another unit PF after a predetermined time. No information is transmitted at all. In this case, each unit PF operates the other unit PF failure diagnosis instruction transmission function unit 12 to transmit request information to the unit PF to which the management right has been given to perform failure diagnosis.

Since the unit PF that has caused the serious failure does not respond at all, each unit PF determines that the unit PF to which the management right has been granted has a serious failure and newly enters the management right. Is selected (SA7).

The selection procedure at this time is as follows.
That is, when each unit PF operates the failure diagnosis function unit 5 and the operation status diagnosis function unit 7 and the load is normal and lighter than the preset load reference, the operation status notification function unit 8 sends the load information to all the units PF. To broadcast.

Upon receiving the load information from each unit PF, the unit PF selects its own and other load information.
And recognizes that when the unit PF is in the lightest load state, it becomes a target to be newly given the management right, and activates the management function (SA8).

As described above, in the present embodiment, when the unit PF to which the management right is given has a trouble due to an overload state or a failure, the management is performed based on the failure information and the load information from each unit PF. A new management right is transferred to the normal and lightest load unit PF between the unit PF to which the right has been granted or another unit PF, and a system cover in a case where the management control function cannot operate normally. The reliability of the management control function has been improved.

Next, the details of the method for granting the management right of the above-mentioned distributed management function agreement type platform will be described. FIG. 10 illustrates a case where the management right is abandoned when the load status of the unit PF to which the management right is assigned is overloaded, and a unit PF to which the management right is to be newly assigned is selected.
FIG. 9 is a diagram showing a flow of a series of processes for operating a management function on the selected unit PF.

The unit PF to which the management right is given periodically diagnoses its own load status by the operation status diagnosis function unit 7 (SB1). Next, it is determined whether the load value obtained as a result of the diagnosis is equal to or less than a predetermined reference value, for example, 0.5 (SB2). As a result of this determination, if the obtained load value is 0.5 or less and there is no overload, the management function is continuously operated, and the normal operation is performed until the next diagnosis time (S
B3).

The load value obtained by the diagnosis is 0.5
Therefore, when there is an overload condition, there is a risk of affecting the entire system. Therefore, the management right is transferred to another normal and lightly loaded unit PF.

As the procedure, the unit PF to which the management right in the overload state is given broadcasts rescue information for relinquishing the management right to all the units PF by the operation status reporting function unit 8 (SB4). Each unit PF that has received this is diagnosed by the operation status diagnostic function unit 7 of its own load state (SB
5) It is determined whether the load value obtained as a result of the diagnosis is equal to or less than a predetermined reference value, for example, 0.5 (SB6). As a result of this determination, if the obtained load value exceeds 0.5 and is overloaded, the own business process is continuously performed (SB
7).

The load value obtained by the diagnosis is 0.5
In the following, when the own load state is lighter than a preset load reference, the own load information is transmitted to the operation status reporting function unit 8.
Is returned to the unit PF to which the management right is assigned (SB8).

Thereafter, the unit PF to which the management right has been assigned sorts the returned information (SB9), and selects the unit PF in the lightest load state from the information by the selection function (SB10). ), The unit PF to which the management right is newly given is determined (SB11). Then
The unit PF to which the management right has been granted transfers the management right to the determined unit PF (SB12), and notifies the transfer of the management right to all the units PF (SB13), and the management function Is stopped, and only normal business processing is performed (SB14). Also, the unit PF to which the management right is newly given operates the management function (S
B15), and the process proceeds to SB1.

In this embodiment, the load reference value for judging an overload is set to 0.5 as an example. However, the present invention is not limited to this.

FIG. 11 shows to all the units PF that the unit PF that has held the management right has delegated the management right to the unit PF that has been newly given the management right in the same situation as in FIG. FIG. 4 shows a flowchart of the processing when it is performed. FIG.
In the processing flow shown in (1), the unit PF to which the management right has been granted is notified to all the units PF that the management right has been moved, but here, the management right is newly assigned. The unit PF performs the operation. However, in the figure, the process of selecting the unit PF to which the management right is newly given is omitted because the broken line area in FIG. 9 is common.

That is, the unit P to which the management right is assigned
In F, when the unit PF to which the management right is newly given is determined (SB11), the unit PF to which the management right is given transfers the management right to the determined unit PF (SB16) and performs the management function. Is stopped, and only normal business processing is performed (SB17).

The unit PF newly given the management right operates the management function (SB18), and notifies the operation status reporting function unit 8 that all the unit PFs have been given the management right. (SB19), the SB
Move to the processing of 1.

Next, the unit P to which the management right is given
A detailed method of selecting a unit PF to which a management right is newly given when a serious failure occurs in F and giving the management right to the unit PF will be described.

FIGS. 12 and 13 are diagrams showing the flow of processing for selecting a unit PF to which a new management right is to be given when a serious failure has occurred in the unit PF to which the management right has been given.

The unit PF to which the management right has been assigned is:
The failure diagnosis function unit 5 is periodically operated to diagnose its own failure state (SC1, SC2). As a result, when it is determined that a minor failure has occurred, the unit PF to which the management right has been granted has the information (relief information) for relinquishing the management right to all units PF on the network. Broadcast by the reporting function (SC3).

Each of the units PF receiving this diagnoses their own failure status and load status by the failure diagnosis function unit 5 and the operation status diagnosis function unit 7 (SC4).
It is determined whether the normal and obtained load value is equal to or less than a predetermined load reference value, for example, 0.5 (SC5).

As a result of this determination, if the load value exceeds 0.5, the own business process is continued (SC6). If the load value is 0.5 or less, the candidate for the management right acquisition and the own load information are transmitted. It is transmitted to the unit PF to which the management right has been given (SC7).

Thereafter, the unit PF to which the management right has been given
Sorts the received load information of each unit PF (SC
8) The selection function unit 10 compares all the load information (SC9), selects the unit PF in the lightest load state (SC10), and delegates the management right to the unit PF (SC11), and thereafter manages the unit PF. Stop the operation of the function (SC
12) Only the operation of the normal unit PF is performed.

The unit PF newly given the management right activates the management function (SC13), and notifies all the units PF on the network that the management right has been given by the operation status reporting function unit 8. (SC1
4) The process proceeds to SC1.

If the result of the diagnosis in SC1 shows that the unit PF to which the management right has been assigned has not failed, the management function is continued (SC15).

On the other hand, if the unit PF to which the management right has been assigned cannot perform any operation due to a serious failure, the unit PF to which the management right has been assigned is transferred to another unit PF after a predetermined time. No information is transmitted at all.

In this case, each unit PF is connected to another unit P
By operating the F failure diagnosis instruction transmission function unit 12, the request information is transmitted to the unit PF to which the management right has been given to perform failure diagnosis (SC16).

As a result of this diagnosis (SC17), if no failure has occurred in the unit PF to which the management right has been granted, the process proceeds to SC15, and if a failure has occurred, the process proceeds to SC3. .

The unit P to which the management right is assigned
When there is no response from F, each unit PF determines that the unit PF to which the management right is granted has a serious failure, and each unit PF operates the operation status diagnosis function unit 7 to determine the load status. Diagnose (SC18).

Next, it is determined whether the load value obtained as a result of this diagnosis is equal to or less than a predetermined load reference value, for example, 0.5 (SC19). The business process is continued (SC20). When the load value is equal to or less than 0.5, the self-load information is broadcast to all the units PF via the operation status reporting function unit 8 and the load information of the other units PF is received (SC2).
1).

Further, each unit PF compares its own load information with the load information of other units PF by the selection function unit 10 (SC22), and determines whether or not its own load state is the lightest load state. A determination is made (SC23).

If the result of this determination is that the self load state is not the lightest load state, the processing shifts to the process of SC20, and if the self load state is the lightest load state, the flow shifts to the process of SC13, and a new management right is granted. Recognize what is to be done and activate management functions.

In this embodiment, the load reference value for judging overload is set to 0.5 as an example. However, the present invention is not limited to this.

As described above, since each unit PF holds the management control function unit 30, the unit PF having the lightest load in the business process performs the management control function, and thus the load situation that changes every moment can be reduced. The corresponding management function unit can be rotated. Thereby, even if the unit PF to which the management right is granted is hindered by an overload state or a minor failure, the failure diagnosis function and the operation state diagnosis function provided inside the unit PF can recognize the other situation,
The unit PF to which the management right has been given based on that can select a unit PF to which the management right can be newly given.

The unit P to which the management right is assigned
Even if a failure occurs in F due to a serious failure, the other unit PF failure diagnosis instruction transmission function unit 12 provided in the unit PF control function unit 32 is given a management right not to transmit any communication even after a certain period of time. The serious failure of the unit PF can be found, and the other units PF can recognize their own status by the failure diagnosis function unit 5 and the operation status diagnosis function unit 7 provided inside the unit PF. Since the unit PF to which the management right can be assigned can be selected, the management control unit 30 is made redundant, and the reliability of the entire system can be improved.

[0062]

As described above, according to the first aspect of the present invention, even if a serious failure occurs in the unit platform to which the management right is given, the serious failure can be found, Other unit platforms can perform failure diagnosis and load state diagnosis inside themselves and recognize their own state, and based on this, can select a unit PF to which a new management right can be assigned, so that the management control function can be performed normally. It is possible to cover the system when it cannot operate, and to enhance the reliability of the management control function.

[Brief description of the drawings]

FIG. 1 is a configuration diagram showing a unit platform according to an embodiment of the present invention.

FIG. 2 is a diagram illustrating a relationship between a database and an application program in a conventional example.

FIG. 3 is a diagram illustrating a relationship between a task module and an application program in a conventional example.

FIG. 4 is a configuration diagram showing a unit platform in a conventional example.

FIG. 5 is a diagram illustrating a system switcher in a conventional example.

FIG. 6 is a configuration diagram showing a unit platform in one embodiment of the present invention.

FIG. 7 is a diagram illustrating a data transfer network according to an embodiment of the present invention.

FIG. 8 is a diagram showing connection of unit platforms in one embodiment of the present invention.

FIG. 9 is a flowchart illustrating an outline of a management right grant method according to an embodiment of the present invention.

FIG. 10 is a diagram showing a flow of a management right granting process when a load status of a unit platform to which a management right is given is overloaded in one embodiment of the present invention.

FIG. 11 is a diagram showing a flow of a management right granting process when a load status of a unit platform to which a management right is given is overloaded in one embodiment of the present invention.

FIG. 12 is a diagram showing a flow of a management right granting process when a serious failure occurs in a unit platform to which a management right is given in one embodiment of the present invention.

FIG. 13 is a diagram showing a flow of a management right granting process when a serious failure occurs in a unit platform to which a management right is given in one embodiment of the present invention.

[Explanation of symbols]

1 ... AP interfacer, 2 ... M interfacer, 3 ...
DB interfacer, 4 accessor, 5 failure diagnosis function unit, 6 failure notification function unit, 7 operation status diagnosis function unit,
8: operation status report function unit, 9: operation status information storage unit, 1
0: selection function unit, 11: connection switching function unit, 12: other unit PF operation status diagnosis instruction transmission function unit, 20: processing unit,
30: management control unit, 31: unit PF management function unit, 3
2 ... unit PF control function _{unit, AP 1, AP 2, AP} 3
... application _{program, M 1, M 2, M} 3 ... task _{module, PF 1, PF 2, PF} 3 ... unit platform.

Continuation of the front page (56) References Takashi Inoue, Norio Kashima, Construction method of subscriber line Ops platform aiming at load distribution, Proceedings of the 1994 IEICE Autumn Conference, Japan, Japan Child Information and Communication Society, September 5, 1994, Communication-2, P.S. 357 Takashi Inoue and Nobuo Kashima, Highly Reliable Configuration of Subscriber Line Ops Platform, Proc. Of the 1994 IEICE Spring Conference, Information Processing Society of Japan, March 10, 1994, 4, p. 4-56 (58) Field surveyed (Int. Cl. ⁷ , DB name) G06F 15/16-15/177 G06F 13/00 H04L 12/24

Claims (1)

(57) [Claims] 1. A database group, a task module group that performs a specific task, an application program group that performs a series of tasks using the database group and the task module group, and the database group, a task module group, and an application program group the a management granting process of the distributed management functions agreement type platform in a distributed processing system configured with a plurality of unit platform is a coupling portion for coupling via a network, the unit platform have been granted administrative rights And periodically communicates with other unit platforms on the network for load distribution and operation status diagnosis. On the other hand, when the communication is interrupted even after a preset time, the unit platform without the management right performs its own failure diagnosis on the unit platform with the management right. After the command is transmitted, when there is no response from the unit platform to which the management right is granted, each unit platform to which the management right is not granted has a priority or a busy status of the application program to be processed by itself. Diagnosis of the operation status and load status such as the average value of the task module and the operation time, and notifies the information to other unit platforms on the network. Select the lightly loaded unit platform from the notified information A new management right can be assigned to a unit platform, and the management right is assigned to the unit platform.