JPH0488533A - Distributed information processing system - Google Patents

Abstract

PURPOSE:To decide the matching of entire system and to update system information by referring to the system information managed by summarizing when the change request of the system information is issued from an arbitrary processor. CONSTITUTION:When the system information is updated at a processor 12a, the update command of the system information inputted from a work station 14a is accepted with a command input processing module 21, and update information is delivered to a matching decision module 22. The matching decision module 22 decides whether or not contradiction exists based on information held in its own processor. When no problem exists, the transmission request of the update information is outputted to a message transmission module 23. The message transmission module 23 receiving the transmission request converts the update information to an update information message, and it is sent out to a transmission line via a transmission control module 24.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a data processing system in which a plurality of processors are connected by transmission paths and communicate with each other while processing and utilizing data, and in particular, , relates to a distributed information processing method in which information regarding the entire system (system information) generated by each processor is distributed and updated by each processor while maintaining consistency among the processors.

[Prior Art] Conventionally, in a data processing system in which a plurality of processors are connected by a communication network and process data using a common database, it is difficult to access any other processor from one processor in the network. It is known that a "processor name" is defined as identification information that can uniquely identify each processor (for example, "Hitachi Office Processor HITACL").
-70 Series MIO87/ES System Manual” (Reference 1)). This processor name is an example of system information, and in this conventional technology, each processor in the system (inside the network) mutually defines its own processor name and the name of the processor to be accessed. There is.

Also, for example, Japanese Patent Application Laid-Open No. 62-208159 (Document 2
), system configuration information (for example, data processing systems and their This includes data processing system configuration information such as device type, presence or absence of communication means, communication protocol, and database system configuration information such as data set type and data set name. Instead of having each processor independently manage system configuration information, it is known to provide a common storage device that can be accessed by each processor and store the system configuration information therein.

Also, for example, Japanese Patent Application Laid-Open No. 63-85854 (Document 3)
As described in , in a distributed processing system in which multiple processors share (distribute) and execute a series of processes to accomplish one job, each processor changes its own program configuration. A system is known that is provided with a means for notifying all other processors of a message to that effect when the system has received the notification, and a means for recognizing the application program configuration of the entire system by using this message when this notification is received. ing.

[Problems to be Solved by the Invention] In the prior art as in Document 1 above, no consideration is given to unified management of all processors in a network system, and therefore each processor has its own processor name defined individually. In order to ensure that the names of the processors to be accessed and the names of the other processors to be accessed are consistent and unique, each processor must also know all the information regarding the other processors. In addition, the definition of processor name,
No consideration was given to checking the consistency of the entire system at the time of update, and there were problems in that definitional inconsistencies were likely to occur, and even if inconsistencies occurred, they could not be sufficiently checked.

Furthermore, in the conventional technology such as the above-mentioned document 2, when the system configuration information is changed, this system configuration information is stored in a common storage device in order to avoid conflicts in the system configuration information between each processor in the system. However, no consideration is given to providing a specific processor (management processor) that aggregates and manages system information regarding all three processors in the system. Therefore, in this conventional technology, when changing a common database and its system configuration information, the processor that wishes to make the change must send an update permission request to all other processors each time. The problem is that the system configuration information of the common database and common storage device is changed when permission is obtained from the processors of the system, and the procedure to complete the change is cumbersome.

In addition, the processing system of Document 3 has a function of sending a program configuration change message to each processor in the system, and a function of receiving the program change message and transmitting the new program configuration to the entire system based on the received contents. It is necessary to provide a recognition function (management function) for all processors in order to maintain the consistency of the entire system, which would complicate the configuration and increase costs. It turns out.

Note that in order to prevent conflicts between processors, it is possible to specify the processor that can perform updates and prevent other processors from performing updates, but this would prevent any processor from performing updates. It is inconvenient because it cannot be done.

Therefore, an object of the present invention is to solve the above-mentioned problems of the prior art, to enable system information to be freely updated from any processor in the processing system, and to update the system information from each processor that is the update source. Even if the processor does not know the information of the entire system, updates can be made while always maintaining the integrity of the system information without causing any inconsistencies, and updates can be made without the need to obtain permission from all processors in the system. An object of the present invention is to provide a distributed information processing method that can perform the following operations and does not complicate the procedure until the update is completed.

[Means for Solving the Problems] In order to achieve the above object, the distributed information processing method in the data processing system of the present invention uses one or more processors that aggregate and manage all system information generated by each processor in the system. The management processor of the computer is predetermined. When this management processor receives a system information update request from any processor, it refers to the system information that is centrally managed, and determines the system information update information (update content) generated by the requesting processor. The system is configured to determine the consistency of the system as a whole, and when it is determined that there is consistency, update the centrally managed system information with this update information.

In addition, the result of this consistency judgment is the request source (update source)
It is also sent to the processor where the necessary updates are made.

[Effect] The effect based on the above configuration will be explained.

According to the present invention, when any processor attempts to update system information, the update content (update information) is
The information is sent to a pre-specified management processor. This management processor collects, stores, and manages all system information generated by each processor in the system. The updated information received by the management processor is compared with system information previously collected and stored in the management processor to determine its consistency. Here, if it is determined that there is consistency, the system information that has been collected and stored in the management processor until then is updated with this new update information. Therefore, the management processor always has the latest system information.

The above judgment result is also sent to the update information generation source (update request source) processor, and if there is consistency (there is no contradiction in all systems), the system information of the update information generation source is also updated. It will be done.

In this way, system information can be freely updated from any processor while maintaining consistency without causing any inconsistency in the entire system.

In this case, each processor that has generated update information can simply send this update information to the management processor, without knowing the information about the entire system, and the management processor can determine the integrity of the entire system based on the latest information. Since it tells you the result, you can update without creating any inconsistencies.

Note that the management processor has been identified in advance (before updating), so by notifying each processor in advance, or each update source processor can send update information to all processors including the management processor. Notify the
By having only the management processor notify the update information generation source processor of the consistency determination result, users of each update information generation source processor can obtain the required system information without being aware of which processor is the management processor. Updates can be made.

[Example code] Embodiments of the present invention will be described below with reference to the drawings.

(Embodiment 1) FIG. 1(a) is an overall configuration diagram of a system to which the present invention is applied. In FIG. 1(a), 12a.

12b and 12c are processors, respectively, which are connected to the transmission path 1 via transceivers 13a, 13b, and 13c.
They are connected by 1. Each processor has a workstation 14a, 14b, 14c and a disk device 15a, 15b, 15C, respectively. Note that although this embodiment shows a case where a bus-type transmission line is applied, the present invention is not limited to this, and any type of transmission line can be used.

FIG. 1(b) is a block diagram of a program that runs on each processor (12a, 12b, 12c) to implement the present invention, and a diagram showing the flow of processing between blocks.

Now, when the processor (12a) updates system information, the command input processing module 21 receives a system information update command input from the workstation (14a), and then passes the updated information to the consistency determination module 22. The consistency determination module 22 determines whether a contradiction occurs based on information held within its own processor.

If there is no problem, a request to send update information is issued to the message sending module 23. The message transmission module 23 that receives the transmission request converts the update information into an update notification message and sends it to the transmission path 11 via the transmission path control module 24.

On the other hand, 12 processors manage all system information.
c, the processor 12c transmits the update notification message sent through the transmission path to the transmission path control module 24.
The message is captured into the processor and passed to the message receiving module 25.

The message receiving module 25 extracts update information from the received update notification message and passes it to the consistency check module 22. The consistency check module 22 determines the consistency of the transmitted update information based on all system information to see if there are any inconsistencies. If there is no problem,
Update the system information by the update module 26,
A request to send an update completion message is issued to the message sending module 23. If it is determined that the update information causes a contradiction, the system information is not updated and the message sending module 2 requests to send a message that cannot be updated.
Submit on 3rd.

On the other hand, in the system information update ratio processor 2a,
An update result (update completion message or update not possible message) is received from the processor 12c, and the update module 26 updates the system information only when the update completion message is received.

Here, in order to explain the processing method of the present invention in detail, 2.
System information targeted by the present invention will be explained. The system information targeted by the present invention is not particularly limited other than information to be managed as a whole system, but here, as an example, a virtual file name (i.e., a virtual file name that points to multiple different files at the same time) is used. In addition to the names given to individual files that exist as files, we will consider names given to a collection of multiple files (hereinafter referred to as "generic file names").

Generic file names shall be unique throughout the system. All files that exist as individual entities indicated by one generic file name (hereinafter referred to as "element files") have the same file attributes, but their physical location on the system may be on the same processor or on different processors. But it's okay. In addition, there must be no duplication of element files indicated by different generic file names. The number of element files indicated by one generic file name (hereinafter referred to as multiplicity) may be less than or equal to a predetermined value.

FIG. 2 is an example of a generic file name management table 200. For one generic file name 201, the processor name 204. where the element file 203 exists. The element file name 205° and element file attributes 206 are managed. In Figure 2, the multiplicity is 202.
Although the case where the value is "2" is shown, it may be any value. This management table is provided on the disk device for each processor, and is also loaded onto the memory when the processor starts up. Note that the processor that manages generic file names (
Management processors) have entries for generic file names related to all processors in the system, but processors that do not manage generic file names (general processors) only have entries for generic file names related to their own processor. will have the following.

Referring to FIG. 3(a), the processing method in the update source processor when updating the definition contents of a generic file name, which is system information, will be explained in detail.

Step 3o1) of extracting the update information from the update command input from the workstation, the consistency determination module 22 checks for errors in the update information based on the generic file name management table 200 possessed by the processor. (Step 302). The contents of the check include whether the multiplicity is less than a predetermined value;
Check whether all file attributes match, whether there are any duplications with other element file names, etc. (However, consistency with other processors cannot be checked yet). Check whether there is an error in the above error check (step 303), and if there is no error, send the update information as an update notification message to the processor (management processor) that manages all generic file names. (Step 304). If there is an error in step 303, error processing such as displaying an error message on the workstation is performed (step 305), and no update notification message is sent.

Then the processor that manages all generic file names (
The processing method in the management processor will be explained in detail with reference to FIG. 3(b).

As described above, an update notification message is sent to the processor from the processor that updated the definition of the generic file name (the update source processor).

As a result of the message type determination (step 312), the message is an update notification message, so the update information is extracted and the consistency determination module 22 performs the same error check as above (step 313). .

At this point, checks are also made to see if there is consistency with other processors, whether the name is unique, etc. Check whether there is an error (step 314), and if there is no error,
The generic file name management table 200 on the main memory (not shown) and on the disk device is updated according to the update information (step 31!5), and an update completion message is sent to the update source processor (step 316). If there is an error, the generic file name management table is not updated and
A non-updatable message is sent to the update source processor (step 317).

In this way, the generic file name management table possessed by the processor (management processor) always stores updated information regarding the generic file names of all processors in the system. The error check (step 313) in the processor corresponds to the error check of the entire system.

For convenience, the processing method in the update source processor is shown again in Figure 3 (
This will be explained in detail using b).

In response to the update notification message sent earlier, a message indicating the update result sent from the management processor of the generic file name is received (step 311), and the message type is determined (step 312). If it is an update completion message, update the generic file name management table on the memory (not shown) and on the disk device (step 318).
, if the message cannot be updated, the generic file name management table is not updated and error handling is performed (step 31).
9).

In addition, as a process when an update notification message is received in all processors, the update notification message is imported only if the own processor is the management processor data with a generic file name, and if it is not a management processor (in general processors) By adding the process of reading and discarding the update notification message, the update source processor does not need to know which processor is the management processor to which the update notification message should be sent, and only needs to broadcast it to all processors.

In addition, since each processor can update only a part of the system information (the part related to its own processor) while maintaining the integrity of the system as a whole, it is possible to update only a part of the system information (the part related to its own processor) while maintaining the integrity of the system as a whole. It can be updated using only knowledge about the system, and a specific system administrator (an administrator who oversees the entire system) is not required.

Finally, regarding the format of messages flowing on the transmission path,
This will be explained with reference to FIG.

FIG. 4(a) shows a conventionally used message format, with flags (F) 401 and 407 indicating the beginning and end of the message, 402 a destination address (DA) specifying the recipient, and 403 is a source address (SA) indicating a transmission source, and 404 is a serial number (C) necessary for transmission. 405 is an area (DATA) for storing information to be transmitted, and 406 is error detection data (Fe2).

FIG. 4(b) shows the message format used in the method of the present invention. Flags 401, 407, destination address 402, source address 4031 serial number 404゜error detection data 4
06 is similar to FIG. 4(a). In FIG. 4(b), in the area corresponding to the data section 405 in FIG. 4(a),
The message type (update notification, update completion, or update not possible) is set as code 'CC 408 indicating the contents of the data section. In addition, 5I409 indicating the target system information
Set the generic file name as . Further, by setting the command code CD410 and command parameter CP411 of the update command, the contents of the update can be determined.

Here, the above-mentioned broadcast to all processors can be realized by setting a global address to the destination address 402 in FIG. 4(b).

The global address is an address that should be received by all processors connected to the transmission path.

By using the method described above, according to this embodiment, when system information is updated by any processor, updates that would cause inconsistency in the entire system are prevented, and the consistency of system information is always maintained. It becomes possible to hold the

(Example 2) Next, another example (Example 2) of the present invention will be described.

In this embodiment, unlike in the first embodiment described above, the specific processors that manage generic file names are not limited to some extent, but all processors are regarded as management processors, and update notification messages are sent to all processors. This is a broadcast method. Since all update notification messages are sent to all processors, information regarding all generic file names is stored in the generic file name management table held for each processor.

Therefore, in the case of this embodiment, since the error check in the update source processor directly becomes the error check for the entire system, there is no need to process the update completion or ('! update impossible message).

FIG. 5(a) is a processing flowchart in the update source processor in the second embodiment.

In the second embodiment, step 501) of extracting update information using an update command input from a workstation.
As a result of the update information error determination (step 502), if there is no error (step 503), the generic file name management table is updated (step 504), and an update notification message is broadcast (step 505).
. If there is an error, the same process as in Example 1 (step 3) is performed.
05) is performed (step 506). FIG. 5(b) is a processing flowchart of all processors that have received the update notification message. After receiving the update notification message (step 511), it is sufficient to immediately update the generic file name management table (step 512).

According to the second embodiment, all processors can have all the system information, and errors in the entire system can be checked only by checking errors in the update source processor. Therefore, consistency of system information can be maintained more efficiently than in the first embodiment.

In any of the above embodiments, system information can be updated from any distributed processor while maintaining the integrity of the entire system, so from the user's perspective,
There is no need to limit a specific processor such as a system information management processor, and operational constraints on the processor as a system can be eliminated. In other words, although the management processor manages system information related to all two processors, it is not intended to restrict updates of system information by processors other than the management processor, and this update can be freely performed by the processor that generated the update information. It is.

In addition, in the first embodiment, instead of broadcasting the update notification message to all processors, at least the management processor is automatically notified by defining it at the same time when determining the management processor (which one is the management processor). By notifying each processor in advance of whether it is a management processor), the user can update the management processor without having to know which processor is the management processor.

[Effects of the Invention] As described in detail above, according to the present invention, a predetermined management processor always has the latest system information collectively regarding all processors (of the entire system),
After that, the consistency of the system information update information generated by any processor is determined based on this latest aggregated system information, and if it is determined that the system information is consistent, the update source and the management processor Since the system information of the system is updated, there is an effect that the system information can be freely updated from any distributed processor while always maintaining the consistency of the entire system.

Further, each processor can update information only if it knows about its own processor without knowing about other processors or the system information of the entire system.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the overall configuration of a data processing system to which the present invention is applied and an example of a program configuration for realizing the processing method according to the present invention. FIG. FIGS. 3 and 5 are process flowcharts for explaining the operation of the processor according to the present invention, and FIG. 4 is a diagram showing the format of messages flowing on the transmission path. 11...Transmission line, 12 a, 12 b, l
2 c −=processor, 13 a , 13 b
, l3c-transceiver, 14a, 14b,
14c...Workstation, 15a, 15b,
15c...Disk 11.21...Command input processing unit module, 22...Consistency determination unit module, 23...Message transmitting unit module, 24...・Transmission line control module, 25・
...Message receiving module, 26...
...System information update module, 200...
Generic file name management table, 401... Message head identification flag, 402... Sending party designation address, 403... Sending source address, 404...
...Transmission serial number, 406...Error detection data, 407...Message termination identification flag, 4
08. Message type, 409... Generic file name (system information), 410... Command code (
update command), 411...Command parameters. Agent: Patent attorney Takeshi Kenjibe (1 other person) Figure 3 Figure 3 (b) 5 (a) (b) Nihi

Claims (1)

[Claims] 1. In a data processing system in which multiple processors are connected by transmission paths, one or more management processors that aggregate and manage all system information generated by each processor are identified in advance, and any processor When a system information update request is received from the management processor, the management processor determines the consistency of the system as a whole with respect to the system information to be updated based on the system information that is collectively managed by itself, and determines whether the system information is consistent. A distributed information processing system characterized in that, when it is determined that the system information to be updated is used, the system information that is centrally managed by the system itself is updated.