JPH02235147A - Multiple control system for program - Google Patents

Abstract

PURPOSE:To prevent processing efficiency from being dropped by providing a system with an executing information collecting means, and when requested source information exceeds its restriction, temporarily storing the information in an execution request holding means to execute multiple control. CONSTITUTION:A restriction information setting processing means A1 sets up system resource information B5 in the case of executing a module to be controlled. At that time, a module characteristic record forming means A2 stores execution load module performance or the like as module characteristic information B3. When an execution load module executing request is generated, transfer processing A3 executes module characteristic information B1. Collecting information is stored in an executing information means A4 as system resource utilizing information to grasp the using state of the resource. When an execution request is outputted and the request exceeds system restriction information, a using resource capacity deciding means A5 holds the execution load module in an execution requesting means A7 until a margin is obtained. Consequently, multiple control is executed in each characteristic unit of a module and the processing efficiency is prevented from being dropped.

Description

[Detailed description of the invention] [Industrial application field] The present invention relates to a program multiplex control system.

[Conventional technology]

In the conventional program multiplex control method, the multiplicity and priority are set for each execution class, and execution load modules (hereinafter referred to as modules) having various characteristics are freely executed within that execution class. There is. By adopting such a method, for example, users who prioritize online work and carry out regular batch work (including regularly executed technical calculation work) and parallel development work are told that ``peak time We often receive complaints that online services have deteriorated and batch processing is not progressing. When examining the computer systems of these users (hereinafter referred to as systems), they often exhibit characteristics such as those shown in Figure 7. Figure 7 shows C for each time period for six execution classes A to F.
It shows changes in PU usage rate and disk usage rate. In reality, there are many more execution classes, but since it would be too complicated to explain, I have only excerpted the main execution classes. Now, looking at the CPU usage rate for each execution class, the peak for execution class A is from 14:00 to 17:00, for execution classes B and C it is almost 0, for execution class D it is from 15:00 to 18:00, and for execution classes E and F it is from 15:00 to 18:00. The peak time is from 15:00 to 18:00.

Moreover, the time when the online service is poor is 2:03 p.m.
0 minutes to 16:30, which coincides with the above time period.

From the above observation results, it can be seen that jobs with a high CPU usage rate have a negative impact on online services.

[Problem to be solved by the invention]

In the conventional method described above, no matter how much the multiplicity of the execution class is lowered, if only modules with high CPU usage are executed, the online service will deteriorate due to CPU waiting (for example, if the modules with high priority Even if you can use it quickly, you will need time to page in the program). On the other hand, if only I/O-based modules are used, execution will be restricted even when the CPU is under low load, resulting in a drawback that effective use of CPU resources will be impaired.

In addition, it is not possible to judge the CPU usage rate and required memory amount at the end of the execution, which are the characteristics of the executed module, and to stop the execution. If usage is concentrated on the CPU, memory, etc., the modules waiting for a certain amount of resources will be exchanged within the system, which is directly proportional to the degree of multiplicity, resulting in a rapid increase in overhead time.

The purpose of the present invention is to prevent the system from becoming overloaded and reducing processing efficiency by performing multiple program control for each module characteristic. [Means for Solving the Problems] The program multiple control method of the present invention is such that a plurality of batch processing or online processing execution load modules are executed simultaneously and repeatedly under the same environment for a certain period of time or after a certain period of time. In a program multiplex control method in a computer system in which the execution load module is created, a limit information setting means for setting limit values for the entire computer system and for each execution load module characteristic in system resource limit information; module characteristic record creation means for storing the performance, required resources, module characteristics, etc. of the execution load module as module characteristic information after execution of the module; and executing the corresponding module characteristic information when there is a request for execution of the execution load module.・Transfer means to transfer to the requesting module characteristic table, and aggregate information in the transferred information and statuses that are being executed for each module characteristic and the entire system, and use the aggregated information as system resource usage status information. Execution information aggregation means that stores and grasps the resource usage status; and module characteristic information of the execution load module that has been newly requested to be executed is temporarily added to the system resource usage status information, and the result exceeds the system limit information. and an execution request suspension means for suspending the execution load module until there is enough room in the computer system to execute the module. Suppose that [Example] Next, this example will be explained with reference to the drawings. FIG. 1 is a block diagram of one embodiment of the present invention.

Referring to FIG. 1, the embodiment of the present invention includes a restriction information setting process A1, a module characteristic record creation process A2, and a transfer process to transfer the module characteristic information B1 requested for execution to the module characteristic table B2 being executed/requested. A3 and
Executing information aggregation process A4 and usage resource capacity determination process A
5, module execution/updating process A6 based on new module characteristics when the judgment of the usage resource allowance judgment processing A5 is executable, and the case where the judgment of the usage resource allowance judgment processing means A5 is not executable. It consists of execution request pending processing A7. In the limit information setting process A1, system resource limit information B5 for executing the module managed by this method is set.
Set. As shown in Figure 4, the details of the system resource limit information B5 include system-wide limit information indicating the maximum limit that can be used or executed by modules managed by this method, regardless of module characteristics, and module characteristics. It is set per module, uses one table for each module characteristic, and consists of restriction information for each module characteristic indicating the maximum value of the limit that the module can use or execute. Furthermore, the system-wide limit information is the maximum number of executable modules (3e in Figure 3>.Ht large usable memory size 3
It consists of three items: f, maximum available CPU usage rate 3g,
Restriction information by module characteristic is module characteristic code 3.
It consists of four items: a (x), maximum executable module number 3b (x), maximum usable memory size 3c (x), and maximum available CPU usage rate 3d (x).

Referring again to FIG. 1, module characteristic record creation processing A2 includes a method for creating module characteristic information B1 based on modules created by compiling and combining resource programs, and a method for a program administrator to operate existing module characteristic information. There is a way to tune it accordingly.

FIG. 5 shows details of module characteristic information, including module characteristics 5a, module storage library 5b, module name 5c, status status 5d, and required memory size 5f.
, average CPU usage rate 5g. It consists of 8 items: record update count 5h and record update date and time 51. Condition status 5d is unused and unused due to 0.1.2 and 3.9, respectively.
Indicates execution requested, execution in progress, execution pending, and execution completed.
In addition, O is given as the initial value for the state status 5d, average CPU usage rate 5g, and record update count 5h. Transfer processing A3 searches the module characteristic information B1 of the module for which execution has been requested from the module characteristic information file, and sets the performance status of the corresponding module characteristic information B1 to execution requesting (code is 1) and changes the characteristics of the module being executed/requested. Transfer to table B2. Note that the transferred module characteristic information B1 indicates that the module execution is completed and
It is managed in the actual/requested module characteristic table B2 until it is deleted in the update process A6.

In the running information aggregation process 4, running/requesting module characteristic information B3 whose status status is "running" (status status is 2) in the running/requesting module characteristic table B2
The required memory size 5f and the average CPU usage rate 5g are totaled for each module characteristic, and the number of modules whose status status is in execution is defined as the number of executed modules, and the memory size used in system resource usage information B4 is calculated as 4c ( x)
, CPU usage rate 4d(x) and number of execution modules 4b
(x) respectively.

Next, the number of executed modules 4b(x) and the memory size used 4c in the table resource usage status by module characteristics (1) to (n) in the system resource usage status information B4.
(x) and CPU usage rate 4d(x),
Number of execution modules 4e, memory size used 4f and C
The PU usage rate is stored in 4g, which is the amount of resources used in the entire system. In the utilization resource capacity determination processing A5, the status status is pending in the executing/requesting module characteristic information B3, and
Information with an old record update date and time is selected, and it is determined whether it is executable or not based on the conditions shown in FIG.

Referring to FIG. 2, in process C1, the usage status by module characteristic corresponding to the module characteristic of the module for which the execution request was made is searched from the resource usage information by module characteristic (FIG. 4), and the table position is set as X. do. Processing C
2. In process C3 and process C4, a limit check is performed by comparing the number of execution modules, memory size used, and CPU usage rate with their respective maximum values, and when the limit values are not exceeded under all conditions (in the flowchart in Figure 2), When the condition "≧" is satisfied), it is determined that it is executable and processing C5 is performed. Also, if this is not the case, it is determined that it cannot be executed. In process C2, the limit on the number of modules that can be executed simultaneously is determined for the entire system and for each module characteristic. In process C3, the limit on the memory size that can be used is determined for the entire system and for each module characteristic. In process C4, limits on the CPU usage rate that can be used are determined for the entire system and for each module characteristic.

In the update process A6, execution of the requested module is permitted, and the state status of the executed/requested module characteristic information B3 is set to being executed. After the execution of the module is completed, the executing/requesting module characteristic information B3 is deleted, and the amount of resources that have been used and whose use has been restricted is changed to the system resource limit information B3.
Release (reduce) from 4 (Figure 3). This is done by executing the running information aggregation process A4. Next, in order to make the module characteristic information up-to-date based on the module execution results, the following tuning (module characteristic information is averaged for each execution to obtain an optimal value) is performed to update the module characteristic information. a. Average CPU usage rate 5g = (average CPU usage rate 5g x number of record updates 5h, CPU usage rate of the module executed last time) ÷ (number of record updates 5h + 1) b. Number of record updates 5h = number of record updates 5h + 1 In execution request pending processing A7, Executing/requesting module characteristic information B3 is a method for suspending execution when it is determined that there are insufficient resources or the execution environment is under high load.
The state status of is set to Pending. When on hold, after a certain period of time has elapsed, execution information aggregation processing A4 and usage resource allowance judgment processing A5 are executed, and these processes are repeated until usage resource allowance judgment processing A5 determines that it can be executed. Holds the module requested for execution.
Figure 6 sets specific values from Figure 1 and illustrates the actual judgment method in an easy-to-understand manner. The illustration focuses on process A5. As a precondition, it is assumed that restriction information according to the system environment is set in the system resource restriction information B5 in consideration of processing capacity.

In such a system environment, module names a1 and a2
, a3 and a4 are being executed (state status is executing in the executing/requesting module characteristic table B2),
It is assumed that the executing/requesting module characteristic information B3 is aggregated for each module and for the entire system, and is stored in the system resource usage status information B4.

In such a state, the determination of the allowable amount of resources to be used when a new execution request is made for module name b1 (see FIG. 2 for conditions) will be divided into (1) to (4) and illustrated. (1) Find a table in which the module characteristic code Fl (for example, 11) is the same as the module characteristic code in the table of module characteristic restriction information B5, and locate the table at table position (
In the example, find table position 7). (2) Limiting the number of modules that can be executed simultaneously (process C2 in Figure 2) D5<E5+1 (18<4+1 in the example) D
2<E2+1 (3<2+1 in the example) above 2
Since both conditions are false, it can be determined that it is executable. (3) Determining the limit on the Emory size that can be used (processing C3 in Figure 2) D6<E6+F3 (1024KB <756KB+64KB>D in the example)
3<E3+F3 (in the example, 250KB<128KB +64KB) Since both of the above two conditions are false, it can be determined that it is executable. (4) Judging the limit on the CPU usage rate that can be used (
Figure 2 Processing C4) D7<E7+F4 (Example Te?.;k 90%<65
%+20X>D4<E4+F4 (50% in example
(40% + 20%) The limit judgment for the entire system in the upper row is false, so it is possible to execute, but the limit for each module characteristic in the lower row is true, and it is determined that it cannot be executed. Execution of the module is permitted only when all of the restriction judgments in (2), (3), and (4) above are determined to be executable; otherwise, execution is suspended. [Effects of the Invention] By employing the configuration described above, the present invention enables multiple control of programs that match the characteristics of the modules executed by the user. In other words, it is possible to set the amount of system resources (consumption amount) that can be used for each module characteristic, and to limit the amount of resources used by the set value.
It is possible to judge whether or not the module can be executed within the limit values assigned to each module characteristic, and if it is determined that it cannot be executed, the module can be suspended immediately before execution. In addition to suspending excessive execution requests that are submitted to the system, it also allows execution of modules that primarily utilize resources that are not used much within the system, allowing system resources to be used to the maximum. There is an effect that can be done.

[Brief explanation of the drawing]

Fig. 1 is a block diagram of an embodiment of the present invention, Fig. 2 is a flowchart for determining the allowable amount of resources to be used, Fig. 3 is a layout diagram of system resource limit information, and Fig. 4 is a layout of system resource usage information. 5 is a layout diagram of module characteristic information, FIG. 6 is an explanatory diagram of the operation of this embodiment, and FIG. 7 is a diagram of the operating status of the computer system. A1... Restriction information setting processing, A2... Module characteristic record creation processing, A3... Transfer processing, A4...
Executing information aggregation process, A5...Usable resource capacity determination process, A6...Update process, A7...Execution request pending process, B1...Module characteristic information, B2...Module being executed/requested Characteristics table, B3... Executing/requesting module characteristic information, B4... System resource usage status information, B5... System resource limit information. Teigou

Claims (1)

[Scope of Claim] A program multiplex control method in a computer system in which a plurality of batch processing or online processing execution load modules are executed simultaneously and repeatedly under an equivalent environment for a certain period of time or after a certain period of time, comprising: limit information setting means for setting limit values for the entire computer system and for each execution load module characteristic in system resource limit information; and module characteristic record creation means for storing module characteristics and the like as module characteristic information; transport means for transferring the corresponding module characteristic information to an executing/requesting module characteristic table when there is an execution request for the execution load module; Execution information aggregation that stores the aggregated information as system resource usage status information and grasps the resource usage status by aggregating the transferred information and status statuses that are being executed for each module characteristic and the entire system. means for temporarily adding module characteristic information of an execution load module newly requested for execution to the system resource usage status information, and determining that the module cannot be executed when the result exceeds system restriction information; 1. A program multiplex control system, comprising: a resource capacity determination means; and an execution request suspension means for suspending an execution load module until there is enough room in the computer system to execute it.