JP2001202259A - Distribution system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To equally distribute a load in a distribution system. SOLUTION: The distribution system is provided with a start receiving means for receiving a server starting request issued by a client by using an identifier name for a prescribed request, a managing table means for managing a correspondence relation between a desired performing file and at least the identifier name for the request and retrieving the candidate of a performing processor being a processor for performing the processing content of the desired performing file by a processing performing part based on a prescribed retrieval rule, a load monitoring means for monitoring the state of a load to be applied to the throughput of each processor, and a performing processor selecting means for selecting a processor in an optimal load state based on a prescribed selecting rule from one or plural processors retrieved as the candidates of the performing processor by the managing table means to distribute the load of each processor in the case of starting desired server processing.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a distributed system, for example, in a distributed system composed of a plurality of processing devices of various types, which can be applied to a case where the load on the processing capacity of each processing device is distributed. It is.

[0002]

2. Description of the Related Art Generally, an application program providing a certain function is called a "server", and an application program using a server function is called a "client". The function of each server is realized by starting and executing an executable file that provides the function of the server.

In a distributed system composed of a plurality of processing devices connected by a network, the function of a server activated on a certain processing device is controlled by a client activated on an arbitrary processing device on the distributed system. Accessible.

[0004] Therefore, when a plurality of servers are simultaneously executed in parallel in the distributed system, there is a problem how to load each processing device with the processing capacity required for executing the servers.

[0005] As a specific load bearing method, for example, there are the following methods (1) and (2).

(1) An execution file of a server is fixedly associated with a processing device in advance, and when a client requests provision of a function, a server providing the function is always associated with the server. It starts on the processing device.

(2) When a client requests the provision of a function, the processing device to which the load required to execute the executable file of the server providing the function is sequentially changed in the order in which the requests are issued. I do.

In this method (1), the load for executing the server is likely to be concentrated on a certain processing device, but in the method (2), the load is distributed to each processing device in the distributed system. Can be.

[0009]

The hardware and operating system of each processing device connected to the distributed system are not always the same. Therefore, an execution file of a server generated on one processing device may not be able to be started on another processing device. Therefore, the plurality of processing devices connected to the distributed system need to be classified into several types among those sharing the executable file that can be started. This classification is called “processing device type”. That is, an execution file for each processing device type can be generated for each server. In other words, the executable on each server is
There are several minutes of processing equipment types.

Therefore, it is necessary to distribute the load in consideration of the type of the processing device.

When an execution file is specified in a server start request from a client, the server can be started on any processing device of a processing device type corresponding to the execution file.

On the other hand, when the execution file is not specified in the server start request from the client, an arbitrary execution file is selected from a plurality of execution files corresponding to the server, and a processing device type corresponding to the execution file is selected. It can be started on any processing device.

[0013] However, even if an execution file is specified in the start request, if there are a plurality of processing units of the processing unit type corresponding to the execution file, the processing unit for starting the server is selected from the plurality of processing units. The question is how to choose.

For example, it is conceivable that the same processing device is always selected according to the method (1), or the processing devices are simply selected sequentially according to the method (2).

However, in the method of always selecting the same processing device, only the processing device is loaded, which is not preferable from the viewpoint of fair distribution of the load.

In the method of selecting the processing devices in order, the load can be fairly distributed at first glance, but there are the following problems, for example.

FIG. 2 shows the processing devices P1 to P5 belonging to the same processing device type and included in the same distributed system 10 and connected by the network NE1, and the load situation of the server processes SV1 to SV10 in each processing device. I have.

Referring to FIG. 2, first, in response to a server activation request issued from any client at time S1, the processing device P1 starts activation and execution of the server SV1. Thereafter, similarly, at time S2, the processing device P2 starts execution of the server SV2 in response to a server start request issued by any client, and at time S3, the server SV issued by any client In response to the activation request, the processing device P3 starts executing the server SV3. At time S5, the processing device P5 starts executing the server SV5 in response to a server activation request issued from any client.

Up to this point, the assignment of the load by execution of the first server to the five processing devices P1 to P5 has completed one cycle. At time S6, the processing devices are responded to in response to a server activation request issued from any client. P1 is the server SV6
Start and execution, and sequentially to the processing device P2 and the like.
The execution load of the second server is allocated.

However, the time required for processing by any one server varies depending on the specific processing contents of the server and the request from the client during the processing time. In the example of FIG. 2, for example, the execution time of the process SV1 is a relatively long time from the start at the time S1 to the end at the time T1, but for example, the execution time of the process SV3 is started at the time S3. This is a relatively short time until the end at time T3.

For this purpose, for example, the processing apparatus P1 performs two processes (SV) simultaneously over most of the illustrated time width.
1 and SV6), while the load is large, for example, the processing device P3 has 0 processes shared over most of the illustrated time width, and at most one process (SV3 or SV8). And the load is small.

The number of processes simultaneously executed by each processing device is 3
Even in the case described above or when the number of processing devices is more or less than five, such a variation in load becomes a problem.

Therefore, it is required to reduce the variation in the size of the load that each processing unit bears and to distribute the load more evenly.

On the other hand, if the execution file is not specified in the server start request from the client, it is necessary to select not only the processing device but also the execution file. Is a problem.

For example, it is conceivable to select an execution file of a server and a processing device to be started by the following methods (1) to (4).

Always select the same executable and processor.

The same execution file is always selected, but the processor selects the execution file in order.

All the combinations of the execution file and the processing device are considered and selected in order.

However, in the method as well as in the method, the load could not be sufficiently and fairly distributed to each processing device in the system for the same reason as in the case of FIG. 2 described above.

[0030]

In order to solve this problem, the present invention is provided with a processing execution unit for executing the processing contents of a target execution file which provides a function of a target server processing to be started. In a distributed system configured by connecting the processing devices to each other via a network, (1) start receiving means for receiving a server start request issued by a client using a predetermined request identifier name; It manages the correspondence between the target execution file and at least the identifier name for the request, and based on a predetermined search rule, the processing execution unit executes the processing contents of the target execution file and executes the processing contents of the target execution file. Management table means for searching; (3) load monitoring means for monitoring the state of the load on the processing capacity of each processing device; Executing means for selecting a processing apparatus having an optimal load state from one or a plurality of processing apparatuses retrieved as candidates for the execution processing apparatus by a cable means based on a predetermined selection rule; And (5) distributing the load of each of the processing devices when activating the target server process.

[0031]

DESCRIPTION OF THE PREFERRED EMBODIMENTS (A) Embodiment Hereinafter, an embodiment of a distributed system according to the present invention will be described.

The first embodiment shows an example in which a server start request from a client is made in a manner of specifying an execution file, and the second embodiment shows that a server start request from a client does not specify an execution file. An example in the case of being performed in an aspect is shown.

(A-1) Configuration of the First Embodiment FIG. 1 shows a main part of the configuration of the distributed system 20 of the present embodiment. Each of the processing devices 1 and 2 shown in FIG. 1 corresponds to each of the processing devices P1 to P5. Here, it is assumed that the processing apparatuses 1 and 2 are processing apparatuses belonging to the same processing apparatus type.

In FIG. 1, the processing device 1 includes a server starting unit 11, a processing device selection unit 12, a processing device type information management unit 13, and a processing device load monitoring unit 14. The processing device 2 connected to the network NE2 includes a server activation unit 21, a server activation reception unit 22, an execution file information management unit 23, and a client CL1.

The server activating unit 11 includes the processing device 1
The above is a portion for starting a new server. When the server is started here, the load for the start consumes the processing capacity of the processing device 1.

Similarly, the server activation unit 21 is provided with the processing device 2
This is a part for starting a new server, and when the server is started here, the load for that consumes the processing capacity of the processing device 2.

The client CL1 is an application program installed in the processing device 2,
When a start request is issued from the client CL1, the server start accepting unit 22 cooperates with the processing device selecting unit 12,
One of the server starting units 11 and 21 according to the information stored in the execution file information managing unit 23, the processing device type information managing unit 13, and the processing device load monitoring unit 14. is there.

The server start accepting unit 22 sends the client C
L1, a part that directly communicates with the server activation part 21 or 11, and the execution file information management part 23, and a processing equipment selection part 12 that directly communicates with the processing equipment type information management part 13 and the processing equipment load monitoring part 14. It is. Processing device selecting unit 12 in response to a search request from server activation receiving unit 22
Performs a selecting operation of a processing device for selecting one processing device for starting the corresponding server, and upon receiving a result of the selecting operation, the server start receiving unit 22 causes the server starting unit 11 or 2
1 starts the server.

That is, when the processing device selection unit 12 selects the processing device 1, the server activation reception unit 22 activates the server in the server activation unit 11, and when the processing device 2 is selected, the server activation reception unit 22 activates the server. Start

The processing unit type information management unit 13 manages all processing unit names in the distributed system 20 and information on the processing unit types, and retrieves a plurality of processing unit names belonging to a specified processing unit type. It is.

The processing unit type information management unit 13 stores a processing unit type name list 1 having a data structure as shown in FIG.
3L.

In FIG. 3, the processing device type name PT
(1) to PT (N) designate processing device types for all the processing devices existing in the distributed system 20, and processing device names PN1,..., PNk,... PNm,. , N are different natural numbers other than 1) are the processing device names included in the distributed system 10.

The processing device name of the processing device 1 shown in FIG.
N1 and the processing device name of the processing device 2 is PNk.
Assuming that (k = 2), the processing device type name of the processing device type to which the processing devices 1 and 2 belong is shown in FIG.
PT (1).

In this case, in the distributed system 20, FIG.
, A processing device (processing device corresponding to the processing device name PNm, PNn, etc.) belonging to the processing device type such as the processing device type name PT (N) shown in FIG. Will be.

The processing unit load monitoring unit 14 monitors the load state of all the processing units in the system 20.
This is a part for searching for a processing device with the least load among a plurality of specified processing devices.

The data structure managed by the processor load monitoring unit 14 may be, for example, a load value list 14L shown in FIG.

In FIG. 9, the same processing apparatus name P as in FIG.
On the right side of N1 to PNn, real-time load values LV1 to LVn corresponding to the processing device of each processing device name are stored. That is, the load value of the processing device PN1 is LV1,
The load value of the processing device PNk is LVk, the load value of the processing device PNm is LVm, and the load value of the processing device PNn is LVn.

Each of the load values LV1 to LVn indicates the number of servers being simultaneously executed in the corresponding processing device at that time.

Finally, the execution file information management unit 23
Is a part that manages information on the execution file name and the processing device type that can start the execution file, and searches for the processing device type name corresponding to the specified execution file name.

The data structure managed by the execution file information management unit 23 is the processing device type list 23 shown in FIG.
L.

In FIG. 8, a processing device type name PT (1) corresponds to the execution file name EF1, and a processing device type name PT (N) corresponds to the execution file name EFN. The function realized by executing each execution file specified by the execution file names EF1 and EFN is the client CL
Seen from 1

In this embodiment, since the processing device type common to the processing devices 1 and 2 is PT (1), the problem is whether to start the server by the server start unit 11 or the server start unit 21. The execution file name that specifies the execution file of the new server corresponds to the EF1.

The functions of the server starting units 11 and 21, the processing unit selecting unit 12, the processing unit type information managing unit 13, the processing unit load monitoring unit 14, the server starting receiving unit 22, and the executable file information managing unit 23 are as follows. In this embodiment, each of them is 1
Think of it as a server.

Therefore, before starting the execution file EF1 which is focused on where to start, in the state of FIG. 1, the load value LV1 of the processing device 1 (processing device name PN1)
Is 3 (servers 12 to 14), and the load value LVk of the processing device 2 (processing device name PNk) is 2 (servers 22 to 23).
It is.

The operation of this embodiment having the above configuration will be described below with reference to the sequence diagram of FIG. FIG. 4 includes steps S1 to S12.

(A-2) Operation of the First Embodiment The client CL1 sends the server
A request is made to start the server by designating the execution file name (EF1) (S1). It is assumed that the execution file name includes information indicating the location of the entity of the execution file.

The server start receiving unit 22 searches the execution file information management unit 23 for a processing device type corresponding to the input execution file name (S2).

Here, since the execution file name EF1 is specified, the processing device type name PT (1) corresponding to the execution file name EF1 in FIG. 8 is obtained as a search result, and is returned to the server start reception unit 22. (S3).

The server start receiving unit 22 passes the returned processing unit type name PT (1) to the processing unit selection unit 12 and requests a search for the processing unit name for starting the server (S).
4).

As a result of the search, the processing device selection unit 12 obtains a list of processing device names corresponding to the passed processing device type PT (1) (of the processing device name list 13L shown in FIG.
A return of the return processing device name list 13RL corresponding to the range of PN1 to PNk belonging to T (1) is received from the processing device type information management unit 13 (S6).

Next, the processing device selection unit 12 selects the processing device name corresponding to the processing device with the lightest load among the plurality of processing device names (here, PN1 and PNk) in the return processing device name list 13RL. In order to know, the return processing device name list 13RL is supplied to the processing device load monitoring unit 14 to perform a search (S7).

At this time, as described above, since the load value LV1 is 3 and LVk is 2, at this point, the processing device name PNk corresponding to the smaller load value LVk corresponds to the processing device with the lightest load. The processing device name is returned to the processing device selection unit 12 (S8).

If the number of processing device names in the return processing device name list 13RL is three or more, the processing device name designating one of the three or more processing devices with the smallest load is: It will be returned in this step S8.

In response to this return, the processing device selection unit 12
Returns the processing device name PNk to the server accepting unit 22 as information for specifying the processing device to start the execution file EF1 (S9).

The server start request receiving unit 2 receiving this return
2 issues the server start request by inputting the execution file name EF1 to the server start unit 21 of the processing device 2 corresponding to the processing device name PNk (S10).

The server start unit 21 acquires the entity of the execution file specified by the passed execution file name PNk, starts the server on its own processing device, and returns the start result to the server start reception unit 22. (S11).

Finally, the server activation receiving unit 22 returns the activation result to the client CL1 (S12).

(A-3) Effects of the First Embodiment As described above, according to the present embodiment, when a server is started by designating an execution file name, there are a plurality of processing apparatuses having the same processing apparatus type. If it exists, it is possible to select the processing device with the lightest load (that is, the number of servers running simultaneously with the least number of servers) at the time of activation, and to activate it on that processing device.

Thus, the load of each processing device can be more evenly distributed in the entire system.

(B) Second Embodiment Hereinafter, only the points of the present embodiment that are different from the first embodiment will be described. The difference in configuration is mainly limited to the portion related to the above-described execution file information management unit (23), and the difference in operation is mainly the server start reception unit (22), the execution file information management unit (23). , And a portion related to the processing device selection unit (12).

The functions of these components are expanded or added in this embodiment as compared with the first embodiment.

(B-1) Configuration and Operation of the Second Embodiment As shown in FIG.
It has processing devices 1A and 2A, and has the same overall configuration as the distributed system 20 of the first embodiment.

In FIG. 5, the processing device 1A includes a server starting unit 11, a processing device selection unit 12a, a processing device type information management unit 13, and a processing device load monitoring unit 14. The processing device 2A connected to the network NE2 includes a server activation unit 21,
Server start receiving unit 22a and executable file information managing unit 2
3a and a client CL1.

The functions of components 11, 13, 14, 21, CL1, and NE2, which are denoted by the same reference numerals as those in FIG. 1, are exactly the same as those in FIG.

However, in this embodiment, the client CL1 issues a server start request in a mode in which the server name is specified and the execution file name is not specified. Here, the server name is a name given to uniquely identify each server in the distributed system 30.

The execution file information management unit 23a characteristic of this embodiment includes an execution file name-processing device type name list 23aL having a data structure shown in FIG. 6, and the execution file information management unit 23a of the first embodiment. 23, the server name information is also managed, and an execution file name and a processing device type name list corresponding to the server name (return execution file name-processing device type name list (for example, 23aRL))
The ability to search has been added.

In FIG. 6, the processing unit type name PT
The execution file name corresponding to (1) is EF1, and the execution file name corresponding to the processing device type name PT (N) is EFN.
The execution file name corresponding to the processing device type name PT (P) is EFP, and the execution file name corresponding to the processing device type name PT (Q) is EFQ.

The execution file names corresponding to the server name SVX are EF1 to EFN, and the execution file names corresponding to the server name SVY (X and Y are different natural numbers) are EFP to EF.
FQ.

Here, the processing device type names PT (1) to PT (N) and the execution file names EF1 to EFN using the same reference numerals as those of the first embodiment are the identifier names (file names) of the first embodiment. Since the functions realized by executing the execution files specified by the execution file names EF1 and EFN are specified by the common server name SVX, they are the same as viewed from the client CL1. is there.

Similarly, the functions realized by executing the execution files specified by the execution file names EFP and EFQ are specified by the server name SVY and are the same as viewed from the client CL1.

However, since the function of the server specified by the server name SVX is different from the function of the server specified by the server name SVY, the activation request by the client CL1 uses a server name that does not specify an execution file. In the case of the present embodiment to be performed, if, for example, the server name SVX is specified in the activation request, only the executable file names EF1 to EFN corresponding to the server name SVX and the processing device type are valid in the executable file information management unit 23a. Names PT (1) to PT (N).

Further, the processing device selection section 12a of this embodiment
Expands the function of the processing device selection unit 12 of the first embodiment, and executes a processing device 1 that starts a server from a plurality of processing devices of a plurality of processing device types for one server start request. And the processing device type can be searched.

Further, the server start receiving unit 22a can receive a server start from the client CL1 by designating a server name.

FIG. 7 shows an operation sequence of the distributed system 30 of the present embodiment having the above configuration. The sequence in FIG. 7 includes steps S21 to S32.

In FIG. 7, the client CL1 requests the server start receiving unit 22a to start a server by designating a server name (S21). Here, it is assumed that the server name SVX is specified as an example.

The server start receiving unit 22a executes a list of the execution file name and the processing device type name corresponding to the input server name (SVX), that is, the return execution file name-processing device type name list (23aRL)). A search is performed from the file information management unit 23a (S22), and a return of the return execution file name-processing device type name list is received from the execution file information management unit 23a as a search result (S23).

The server start receiving unit 22a holds the information content of the return execution file name-processing device type name list obtained as a search result, and from this list, only a list of the processing device types (for example, the return Executable file name—extracted processing device type name list composed of only processing device type names PT (1) to PT (N) generated by removing executable file names EF1 to EFN from processing device type name list 23aRL) , Processing device selector 12a
To request a search for a processing device that starts the server (S24).

The processing device selection section 12a inputs a plurality of processing device types in the passed list of extracted processing device types (S25), and lists processing device names corresponding to each processing device type (return processing). The device name list) is sequentially searched from the processing device type information management unit 13 (S26).
The return processing device name list obtained as a result of this search is, for example, a list corresponding to the return processing device name list 13RL.

The processing device selection section 12a holds the information content of the returned processing device name list obtained as a search result, and further searches all the processing device type names in the extracted processing device type name list ( S27), a processing device name designating the processing device with the lightest load is searched from all the processing device names included in the plurality of return processing device name lists obtained as a search result (S28).

This search is performed by using the processing device load monitoring unit 14 as in the first embodiment.

Further, the processing device selection unit 12a inputs the processing device name designating the processing device with the lightest load obtained as a search result, together with the processing device type, to the server start reception unit 22a.
a (S29).

Assuming that the list of the processing device types passed to the processing device selection section 12a in step S24 is only the processing device type (processing device type name PT (1)) to which the processing devices 1A and 2A belong. As in the first embodiment, the servers 12a, 1
Since the number (3) of the servers 3 and 14 is smaller than the number (2) of the servers 22a and 23a running on the processor 2A, in step S29, the processor name PNk of the processor 2A and the processor type PT ( 1) will be returned as a search result.

Upon receiving this return, the server start receiving unit 2
2a requests the server activation unit 21 of the processing device 2A to activate the server by passing an executable file name EF1 that can be activated on the processing device 2A (S30).

The server activation section 21 acquires the entity of the execution file having the passed execution file name EF1, activates the server on its own processing device 2A, and returns the activation result to the server activation reception section 22a (S31). ). The server activation receiving unit 22a further transmits the activation result to the client CL.
1 (S32).

(B-2) Effects of the Second Embodiment As described above, according to the present embodiment, when a server is started by designating a server name without designating an executable file name, Even if there are a plurality of execution files that provide functions and there are a plurality of processing devices that can start each execution file,
It is possible to select a processing device with a light load (that is, a small number of servers running at the same time) and start it on that processing device.

Thus, the load of each processing device can be more evenly distributed in the entire system.

(C) Other Embodiments In the first and second embodiments, the magnitude of the load shared by each processing device is compared based on the number of activated servers. When there is a large variation in the load per unit time required for starting and executing each server to be started, such as a server having a high real-time property and a server having a low real-time property (in the example of FIG. 2, the servers SV1 and SV6 are executed simultaneously). To be borne by the processing device P1 and the server SV
For example, when the loads borne by the processing device P2 for executing only the number 2 are the same, it may not be appropriate to estimate the load based only on the number of servers. In such a case, the load may be estimated based on the usage rate of the CPU (central processing unit) of each processing device.

In FIG. 1 (FIG. 5), the processing device selection unit 12 (12a), the processing device type information management unit 13,
The processing device load monitoring unit 14 is activated on the processing device 1 (1A), the server activation reception unit 22 (22a), and the execution file information management unit 23 (23a) are activated on the processing device 2 (2A). The function only needs to be activated on any processing device in the system 20 (30). That is, it may be activated in a different combination from the state illustrated on the processing device 1 or 2 (1A or 2A), and may be started on any processing device in the distributed system other than the processing devices 1 and 2 (1A and 2A). May be activated.

Further, the server 12 and the like shown in FIG. 1 and the like exemplarily show the minimum types of servers necessary as a premise in the operation of the first and second embodiments.
Therefore, the present invention does not prevent a server other than the illustrated one from being activated in each processing device at the time of requesting activation of a new server to be activated by the server activation unit (for example, 11 or 12). .

In the first and second embodiments, each component is one server, but in the present invention,
A plurality of components, for example, a processing device selection unit, a processing type information management unit, and a processing device load monitoring unit may be considered as one server, and the servers 12 to 14, 22 to 2 may be considered as one server.
It is also possible to consider the functions of the three five servers as four or less servers or six or more servers.

The first embodiment corresponds only to the case where a server start request from a client is made in a mode of specifying an executable file, and the second embodiment corresponds to a case where a server start request from a client specifies an executable file. Although only the case where the specification is not specified is dealt with, the function of the first embodiment and the function of the second embodiment are combined so that the activation request issued from the client does not specify the case where the execution file is specified. A system configuration that can cope with both cases may be adopted.

Further, in the first and second embodiments, the function block for requesting the server to start is added to the client CL1.
1 and 5, the client CL1 exists in the processing devices 2 and 2A. However, the client CL1 may be an application pro program activated on an arbitrary processing device in the system. It may be considered a person such as a system administrator existing outside the system. In the latter case, it is assumed that the server start reception unit has an interface for the outside of the system.

Each processing device may be a workstation or a PC (personal computer), and workstations and PCs may coexist in the same distributed system.

As described above, when there is a large variation in the original processing capacity of each processing device in the system, at the time of the start request (this is almost equal to the start time), the processing device with the smallest load is simply obtained. Instead of selecting the processing device, it is also possible to select a processing device having the largest margin of processing capacity.

Further, the processing device name in the first and second embodiments may be a character string name or an IP address as long as the processing device can be uniquely identified.

In the first and second embodiments, the present invention is realized mainly by software. However, the present invention can be realized mainly by hardware.

[0107]

As described above, according to the present invention, at the time of starting the target server processing, the processing apparatus having the optimum load condition is selected as the execution processing apparatus. On the other hand, it is possible to distribute the load more fairly than before.

[Brief description of the drawings]

FIG. 1 is a schematic diagram illustrating a configuration of a distributed system according to a first embodiment.

FIG. 2 is a diagram illustrating the operation of a conventional distributed system.

FIG. 3 is a schematic diagram illustrating a configuration of a processing device name list used in the first and second embodiments.

FIG. 4 is an operation sequence showing an operation of the first embodiment.

FIG. 5 is a schematic diagram illustrating a configuration of a distributed system according to a second embodiment.

FIG. 6 is a schematic diagram illustrating a configuration of an execution file name-processing device type name list used in the second embodiment.

FIG. 7 is an operation sequence of the operation of the second embodiment.

FIG. 8 is a schematic diagram illustrating a configuration of a processing device type name list used in the first embodiment.

FIG. 9 is a schematic diagram showing a configuration of a load value list used in the first and second embodiments.

[Explanation of symbols]

1, 2, 1A, 2A ... processing apparatus, 10, 20, 30 ... distributed system, 11, 21 ... server starting unit, 12, 12a
... Processing device selection unit, 13 ... Processing device type information management unit
14: processing device load monitoring unit, 22, 22a: server start reception unit, 23, 23a: execution file information management unit.

Continued on the front page F term (reference) 5B045 AA00 BB02 BB12 BB28 BB48 BB49 GG02 GG06 5B089 GA11 JA11 JB22 KA06 KB06 KC15 KC23 KC44 MA03 MA07 5B098 AA10 GA01 GC08 GC16 GD02 GD14

Claims (3)

[Claims] 1. A processing system comprising a processing execution unit for executing a processing content of a target execution file for providing a function of a target server processing to be started, the processing units being connected to each other via a network. In a distributed system, a start receiving unit that receives a server start request issued by a client using a predetermined request identifier name, and manages a correspondence relationship between the target execution file and at least the request identifier name, and uses a predetermined search rule. Management table means for searching for a candidate of an execution processing device which is a processing device for executing the processing content of a target execution file in the processing execution unit, and monitoring a state of a load on a processing capacity of each processing device based on One or more processing devices searched as a candidate for the execution processing device by the load monitoring means and the management table means. And an execution processing device selecting means for selecting a processing device having the optimum load condition based on a predetermined selection rule, when starting the target server process, the load of each of the processing devices is reduced. A distributed system characterized by being distributed. 2. The distributed system according to claim 1, wherein an execution file name for uniquely specifying the target execution file is used as the request identifier name, and the management table means is hardware or software. A correspondence relationship between a processing device group name that uniquely specifies a processing device group composed of one or a plurality of processing devices that permits execution of the same execution file in the processing execution unit, and the execution file name A first element table unit that searches for a processing device group to which the executable file name belongs from the executable file name, a processing device name that uniquely specifies each of the processing devices, and a processing device group name. A second relationship for managing one or more processing device names as candidates for the execution processing device from the processing device group name. Distributed system characterized by comprising a prime table portion. 3. The distributed system according to claim 1, wherein a server processing name that uniquely specifies a function of the target server processing is used as the request identifier name, and the management table means is configured as hardware, software, or the like. Is a processing device group name that uniquely specifies a processing device group composed of one or more processing devices that permits execution of the same execution file in the process execution unit; and the execution file name; A third element table unit that manages a correspondence relationship with the server processing name and searches the server processing name for an executable file name and a processing device group name belonging to the server processing name; Manages the correspondence between the processing device name to be specified and the processing device group name, and as a candidate for the execution processing device from the processing device group name. Distributed system characterized in that it comprises a second element table section for retrieving one or more processors names.