JP2001195271A - Method and device for generating and deciding distributed object server thread - Google Patents

Abstract

PROBLEM TO BE SOLVED: To process a request from a client by validly utilizing the resources of a server by including the situation of a server other than its own distributed object server thread, and to validly distribute the server resources according to the significance of plural distributed object server processes which are present in the server in a method for generating and deciding a distributed object server thread. SOLUTION: This device is provided with a function for obtaining and managing the resource situation of a server in which a distributed object server thread is present, and the resource situation is reflected as the decision factor of the validity/invalidity of the generation of the distributed object server threshold. Thus, it is possible to generate the distributed object server thread by validly using the server resources.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a method for determining whether to create a distributed object server thread that operates in response to a request from a client.

[0002]

2. Description of the Related Art Distributed object technology is standardized by a predetermined organization. In a distributed object server / client system, a method of processing requests from a plurality of clients for each thread in one distributed object server process is widely used. The standard does not mention thread policy, and Objec
t Request Broker (ORB) Implementer implements it with its own thread policy.

FIG. 8 is a flowchart for explaining the operation of a conventional master server process disclosed in Japanese Patent Application Laid-Open No. 7-152590. In Japanese Patent Application Laid-Open No. 7-152590, in a client / server type computer system, the upper limit number of threads generated per process of a server is set to a certain fixed value. If there is a connection request from a new client process and the number of currently connected client processes exceeds a certain value, a slave server process is created. When a connection request is received from a new client process, a thread is created if the number of currently connected client processes does not exceed a certain value. The number of threads that can be created and the number of client process connection requests accepted by the server process are set to a static constant value. Since the conditions differ depending on the amount of the computer resources, the above values are set in consideration of those conditions.

[0004]

As described above, in the prior art, the number of server threads that can be created is set to a constant value.
The value is estimated and set in accordance with conditions such as the number of clients, the scale of work performed by processes and threads, and the amount of computer resources. However, in reality, the resource use status of the server process changes, and the number of server threads that can be generated changes accordingly.

[0005] As a result, for example, if the number of server threads that can be created is set to 3 by estimating the condition under which the distributed object server threads can create up to 3, and the number of created distributed object server threads is 3, the server resources Does not generate a distributed object server thread even if there is room to generate a distributed object server thread. For this reason, there is a case where a client's request is not processed even though it can be processed as a server resource.

Conversely, for example, if the conditions are set so that three distributed object server threads can be created, and the number of distributed object server threads already created is 1,
An attempt is made to create a distributed object server thread even if the server resources cannot afford to create a distributed object server thread. As a result, the resources of the server may be depleted, and the server itself may end abnormally. In order not to deplete server resources, it is necessary to underestimate the number of distributed object server threads that can be created. However, if the estimated number of distributed object server threads that can be created is underestimated, as described above, there may be cases where server resources cannot be processed even though they are client requests that can be processed.

[0007] It is an object of the present invention to provide a method for determining the creation of a distributed object server thread by including the status of a server other than the self-distributed object server thread.
The purpose is to be able to process client requests using server resources effectively. It is another object of the present invention to effectively distribute server resources according to the importance of the distributed object server process when a plurality of distributed object server processes exist in the server.

[0008]

SUMMARY OF THE INVENTION In order to achieve the above object, the present invention relates to a method for determining the creation of a distributed object server thread in a system in which a request from a client is processed by a thread in the distributed object server process. When there is a request from the distributed object server, it is determined whether or not to generate a thread for processing the request based on the state of resources in the distributed object server.

[0009] A method for determining generation of a thread in a distributed object server in a system in which a request from a client is processed by a thread in a distributed object server process, wherein a condition for generating a thread in the distributed object server is defined as a server thread generation availability condition. Setting, obtaining the status of the resources in the distributed object server, and, upon a request from the client, determining whether the obtained status of the resources satisfies the server thread creation enable / disable condition. Deciding whether to generate a thread for processing the request from the client according to the result.

The resource status is, for example, a CPU utilization rate, a memory utilization rate, or the number of running threads. A server thread generation availability condition may be set based on a combination of a plurality of resource statuses.

[0011] The server thread creation feasibility condition defines the number of threads that can be created in accordance with the status of the resource. The step of determining whether to create the thread includes the server thread creation feasibility condition. The number of threads that can be created in accordance with the status of the acquired resources based on the number of threads, and when the number of threads currently created is smaller than the number of threads that can be created, process the request from the client. A thread may be created.

In the distributed object server, a plurality of distributed object server processes are executed, and the step of determining whether or not to create the thread is performed by a distributed object server process other than the self-distributed object server process. Whether or not to create a thread in the self-distributed object server process may be determined based on the resource status including the resource status.

Further, the present invention relates to a distributed object server thread generation determining device in a system in which a request from a client is processed by a thread in a distributed object server process, the server thread being a condition for generating a thread in the distributed object server. Means for setting a creation permission / prohibition condition, means for acquiring a resource status in the distributed object server, and determining whether or not the obtained resource status satisfies the server thread generation permission / inhibition condition when requested by a client. Means for determining whether or not to generate a thread for processing a request from the client according to the determination result.

[0014]

Embodiments of the present invention will be described below in detail with reference to the drawings.

FIG. 1 shows an example of a system configuration in which the present invention is applied to a distributed object client / server model. FIGS. 6 and 7 are flowcharts for explaining the thread generation determination operation in the system of FIG. In particular, FIG. 6A shows a procedure for setting conditions for generating a server thread. FIG. 6B shows a procedure for acquiring the resource status of the server. FIG. 7 shows a procedure for processing a request from a client.

Referring to FIG. 1 and FIG. 6A, a procedure for setting a server thread generation condition in the system of the present embodiment will be described. When a predetermined operation is performed, the server thread generation condition setting process starts (step 20).
1) When it is performed, the distributed object server thread management condition setting unit 12 of FIG.
Alternatively, by reading data described in a file or the like in advance, a server thread generation availability condition is acquired (step 202). The server thread management condition setting unit 12 is a distributed object server thread management unit 1
For 3, the distributed object server thread creation availability condition acquired in step 202 is set (step 20).
3), the process ends (step 204).

Next, a procedure for acquiring a server resource status in the system according to the present embodiment will be described with reference to FIGS. 1 and 6B. The process in FIG. 6B is started in the resource status acquisition unit 14 when the server 1 operates. FIG.
When the process (b) is started (step 205), it is determined whether the distributed object server process 11 is operating (step 206). If the server is in operation, the resource status acquisition unit 14 acquires the usage status of various resources of the server 1 using a function or the like provided by the OS (for example, a vmstat command) (step 207). Next, the resource status acquisition unit 14
The acquired resource status is registered in the resource status management unit 15 (step 208). After step 208, step 206
Return to The resource status acquisition unit 14 determines in steps 207 and 20
8 is repeated at regular intervals to acquire the resource status of the server 1 and register it in the resource status management unit 15. If the distributed object server process is not running in step 206, the process ends without acquiring the resource status (step 209).

Next, a processing procedure of a request from a client in the system of the present embodiment will be described with reference to FIGS. The process in FIG. 7 is started when the server 1 operates. The client 3 sends a request to the server 1 via the network 2. When the process of monitoring the request from the client 3 (Step 210) is started, it is determined whether or not the distributed object server process 11 is operating (Step 211). If it is in operation, it is determined whether there is a request from the client 3 (step 21).
2). If there is no request from the client, step 2
Return to 11. If there is a request from the client, the distributed object server thread management unit 13 acquires the resource status of the server 1 from the resource status management unit 15 (step 2).
13) It is determined whether a thread can be created by comparing the distributed object server thread creation availability condition set by the distributed object server thread management condition setting unit 12 with the acquired resource status (step 214). If a thread can be created, a thread is created and a request from the client 3 is processed (step 21).
5). If a thread cannot be created, an error indicating that processing cannot be performed is returned to the client 3 (step 216). After steps 215 and 216, the process returns to step 211. If the distributed object server process is not running in step 211, the process ends (step 217).

One example of a flow for judging whether or not to create a thread after receiving a client request is as follows. The server 1 that has received the request from the client 3 transmits the request from the client 3 to the corresponding distributed object server process 11. The distributed object server process 11 that has received the request leaves the processing to the distributed object server management thread 111. The distributed object server management thread 111 inquires of the distributed object server thread management unit 13 whether a distributed object server thread can be created. The distributed object server thread management unit 13 determines whether a thread can be created based on the above-described conditions for creating a distributed object server thread and the resource status.

FIG. 2 shows an example of a method for determining whether or not to create a distributed object server thread in step 214 of FIG. The table of FIG. 2A shows that "the CPU utilization is from 0 to 30%" is required as a condition for creating a distributed object server thread. In the figure, ○ indicates that a distributed object server thread can be generated, and × indicates that a distributed object server thread cannot be generated. FIG. 2 (b)
The table indicates that "memory utilization is 0 to 50%" is required as a condition for creating a distributed object server thread.

Combining the conditions of FIG. 2A and FIG. 2B results in a table of FIG. In this way, by combining a plurality of conditions for generating a distributed object server thread, it is possible to flexibly set conditions for effectively using server resources. For example, FIG.
In the example, if the CPU utilization is 30% and the memory utilization is 50%, the distributed object server thread can be created.If the CPU utilization is 10% and the memory utilization is 60%, the distributed object server thread can be created. Server thread creation is not possible.

In the above embodiment, the distributed object server management thread 111 has been described as an example in which the distributed object server thread management unit 13 inquires of the distributed object server thread management unit 13 whether or not a distributed object server thread can be created. When it is detected that the state of the thread generation possibility changes, the state may be notified to the distributed object server management thread 111. In the above description, the client and the server are distributed on the network, but they may be on the same machine.

FIG. 3 shows another example of determining whether or not to create a distributed object server thread in step 214 of FIG. FIG. 3A shows that when the CPU utilization is 0%, the number of distributed object server threads that can be created is 7,
If the U usage rate is greater than 0% and up to 30%, set the number of distributed object server threads that can be created to 5, and the CPU usage rate is 30%
This is a table showing the condition that the number of distributed object server threads that can be created is set to 0 when the number is larger than 0.

FIG. 3B shows that the number of distributed object server thread generations is 7 when the memory utilization is 0%, and that the distributed object server thread generation is performed when the memory utilization is greater than 0% and up to 50%. 11 is a table showing a condition that the possible number is set to 3 and the number of distributed object server threads that can be created is set to 0 when the CPU utilization is greater than 50%.

When the conditions shown in FIGS. 3A and 3B are combined, a table shown in FIG. 3C is obtained. In this way, by combining a plurality of conditions for generating a distributed object server thread, it is possible to flexibly set conditions for effectively using server resources. For example, FIG.
In the example, if the CPU utilization is 30% and the memory utilization is 50%, the number of distributed object server threads that can be created is 5, and if the CPU utilization is 10% and the memory utilization is 60%, The number of distributed object server threads that can be created is 0.

Since the distributed object server management thread 111 generates and manages the distributed object server threads 112, the number of the distributed object server threads 112 can be managed. When receiving a request from the client 3, the distributed object server management thread 111 passes the number of distributed object server threads 112 to the distributed object server thread management unit 13, and in response, the distributed object server thread management unit 13 determines The result of whether or not the created distributed object server thread can be generated may be received.
Further, the distributed object server thread management unit 13 notifies the distributed object server management thread 111 of the number of distributable object server threads that can be generated, and the distributed object server management thread 111 that has received the request from the client 3 exists at the present time. It is also possible to determine whether to create a distributed object server thread by comparing the number of distributed object server threads with the number of generable distributed object server threads received from the distributed object server thread management unit 13.

Next, a second embodiment of the present invention will be described with reference to FIGS. FIG. 4 shows a configuration example of a server according to a second embodiment in which the present invention is applied to a distributed object client-server model. In FIG. 4, the same reference numerals as those in FIG. 1 denote portions that perform the same functions as in the above-described first embodiment, and differences from the first embodiment will be described in detail below.

In FIG. 4, a plurality of distributed object server processes 11 exist in the server 1, and each distributed object server management thread 111 exchanges information with the same distributed object server thread management unit 13. The two distributed object server processes shown in FIG. 4 are called a distributed object server A and a distributed object server B, respectively. In the second embodiment,
One distributed object server thread management unit 13 can generate threads for both two distributed object server processes in consideration of server resources. Here, description will be given focusing on thread generation in the distributed object server B.

FIG. 5 shows the distributed object server B of FIG.
5 shows an example of a condition for determining whether a distributed object server thread can be created or not. The table of FIG. 5A indicates that the condition of the distributed object server thread creation of the distributed object server B is “CPU utilization from 0
Up to 30% ". In the figure, ○ indicates that a distributed object server thread can be generated, and × indicates that a distributed object server thread cannot be generated. The table in FIG. 5B shows that "the number of threads of the distributed object server A is from 0 to 10" is required as a condition that the distributed object server B can generate the distributed object server thread.

When the conditions shown in FIGS. 5A and 5B are combined, a table shown in FIG. 5C is obtained. As described above, when the number of threads of the distributed object server A exceeds 10, the thread of the distributed object server B is not generated, and the distributed object server A can be preferentially processed. Also, by combining a plurality of conditions for enabling or disabling a distributed object server thread, it is possible to flexibly set conditions for effectively using server resources. For example, when the CPU utilization is 30% and the number of threads of the distributed object server A is 10, the distributed object server thread of the distributed object server B can be generated.
If the number of threads is 15, the distributed object server B
Cannot create a distributed object server thread.

The distributed object server management thread 111 of the distributed object server A has a distributed object server thread 112 of the distributed object server A.
Is generated and managed, the number of distributed object server threads 112 in the distributed object server A can be managed. Therefore, it is possible to determine whether a thread of the distributed object server B can be created as follows. First, when receiving a request from the client 3, the distributed object server management thread 111 of the distributed object server B inquires the distributed object server thread management unit 13 whether the distributed object server thread can be created. The distributed object server thread management unit 13 acquires the number of threads of the distributed object server A from the distributed object server management thread 111 of the distributed object server A, and determines whether the distributed object server B can generate the distributed object server thread.

Also, the distributed object server management thread 111 of the distributed object server A notifies the distributed object server thread management unit 13 of the number of threads of the distributed object server A in advance, and every time the number of threads of the distributed object server A changes. Alternatively, the distributed object server thread management unit 13 may be notified of the number of threads of the distributed object server A.

Further, the distributed object server thread management unit 13 notifies the distributed object server management thread 111 of the distributed object server B of the number of distributed object server threads that can be generated by the distributed object server B in advance, and the distributed object server B generates the distributed object server thread. Each time the number of possible distributed object server threads changes, the value may be notified to the distributed object server management thread 111 of the distributed object server B.

Although the second embodiment has been described for the case where there are two distributed object server processes, the same applies to the case where there are three or more distributed object server processes. As an example, a case where a distributed object server A, a distributed object server B, and a distributed object server C exist will be described. The server thread creation permission condition of the distributed object server B is set to "the server thread creation of the distributed object server B is possible when the number of threads of the distributed object server A is 0 to 10." Similarly, the server thread creation availability condition of the distributed object server C is:
"The number of threads of the distributed object server A is from 0 to 3, and the number of threads of the distributed object server B is 0.
When it is 10 to 10, the server thread of the distributed object server C can be created. " Distributed object server C
When a request from a client is received, it is determined whether a distributed object server thread can be created based on the status of the distributed object server A and the distributed object server B.

As described above, when a plurality of distributed object server processes exist in a server, when one of the distributed object server processes receives a request from a client, the status of another distributed object server process (other If the thread creation is determined according to the number of threads created in the distributed object server process, it is possible to perform thread creation with priority among the plurality of distributed object server processes. Server resources can be more effectively allocated and used.

[0036]

As described above, according to the present invention, whether or not to create a distributed object server thread is determined according to the resource status of the server and the status of the distributed object server process of the server itself, and thus the server resources are effectively used. Can be used for

[Brief description of the drawings]

FIG. 1 is a diagram showing an example of a system configuration in which the present invention is applied to a distributed object client / server model.

FIG. 2 is a diagram illustrating an example of a table for explaining a thread generation determination method.

FIG. 3 is a diagram illustrating an example of a table for explaining a thread generation determination method.

FIG. 4 is a diagram illustrating a configuration example of a server according to a second embodiment of the present invention.

FIG. 5 is an example of a table for explaining a thread generation determining method.

FIG. 6 is a flowchart illustrating an operation of setting a thread generation availability condition and acquiring a resource status.

FIG. 7 is a flowchart illustrating an operation of determining whether to create a thread.

FIG. 8 is a flowchart illustrating a server process operation of the related art.

[Explanation of symbols]

Reference Signs List 1 server 2 network 3 client 11 distributed object server process 12 distributed object server thread management condition setting unit 13 distributed object server thread management unit 14 resource status acquisition unit 15 resource status management unit 21 CPU utilization (%)-distributed object server thread generation Availability table 22 CPU utilization rate (%) 23 CPU utilization value 24 Distributed object server thread creation availability 25 Distributed object server thread creation availability value 26 Memory utilization rate (%)-Distributed object server thread creation availability table 27 Memory utilization rate (%) 28 Memory utilization value 29 CPU utilization (%)-Memory utilization (%)-Distributed object server thread creation availability table 31 CPU utilization (%)-Distributed object server thread Generateable Number Table 32 Distributed Object Server Thread Generateable Number 33 Distributed Object Server Thread Generateable Number 34 Memory Usage Rate (%)-Distributed Object Server Thread Generateable Number Table 35 CPU Usage Rate (%)-Memory Usage Rate (%) -Distributed object server thread creation possible table 41 Distributed object server A thread number-Distributed object server thread creation availability table 42 Distributed object server A thread number 43 Distributed object server A thread number 44 CPU utilization (%)-Distributed Object server A
Number of threads of distributed object server thread generation possibility table 111 Distributed object server management thread 112 Distributed object server thread

Claims (5)

[Claims] 1. A distributed object server thread generation determining method in a system in which a request from a client is processed by a thread in a distributed object server process, the method comprising the steps of: A method for determining whether to generate a thread for processing the request based on the request. 2. A distributed object server thread generation determining method in a system in which a request from a client is processed by a thread in a distributed object server process, wherein a server thread generation availability condition, which is a condition for generating a thread in the distributed object server, is set. Setting, obtaining the status of the resources in the distributed object server, and upon receiving a request from the client, determining whether the obtained status of the resources satisfies the server thread creation permission / inhibition condition. Determining whether to generate a thread for processing the request from the client according to the result. 3. The distributed object server thread creation determining method according to claim 2, wherein the server thread creation availability condition defines the number of threads that can be created according to the resource status. The step of deciding whether or not to generate,
Acquiring the number of threads that can be created according to the status of the acquired resource based on the server thread creation availability condition,
A distributed object server thread generation determining method, characterized in that a thread for processing a request from the client is generated when the number of threads currently generated is smaller than the number of threads that can be generated. 4. The distributed object server thread generation determining method according to claim 2, wherein a plurality of distributed object server processes are executed in said distributed object server, and said thread is generated. Determining whether or not
A distributed object server thread for determining whether to create a thread in the self-distributed object server process based on a resource status including a resource status in another distributed object server process other than the self-dispersion object server process. Generation decision method. 5. A distributed object server thread generation determining device in a system in which a request from a client is processed by a thread in a distributed object server process, wherein a server thread generation availability condition, which is a condition for generating a thread in the distributed object server, is set. Means for setting, means for acquiring the resource status in the distributed object server, and, when requested by a client, determining whether the obtained resource status satisfies the server thread creation enable / disable condition, Means for determining whether or not to generate a thread for processing a request from the client according to a result.