JPS63189960A - System performance improving system - Google Patents

Abstract

PURPOSE:To reduce the burden of a person in charge of a maintenance and to rapidly present and execute an improving measure by automatically analyzing performance data and examining a performance improving measure to a certain level. CONSTITUTION:An initial system model table 2 is updated based on the performance improving measure selected by a performance data table 4, an improving measure selecting means 8 in a system model table updating means 10 and an updated system model table 9 is formed. A system performance predictive means 11 inputs the model table 9, calls a performance predictive function 1 to predict a calling performance and informs the result to a system predictive control decision means 7. An improving measure display and executing means 12 displays and execute the improving measure which can be executed during the operation of the system.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a method for improving system performance of an electronic computer system.

(Prior Art) Conventionally, when improving the system performance of a computer system, generally, various performance data (for example, the usage status of the central processing unit) in the system are collected using a software monitor, and the collected performance data is used in various ways. The information was edited in a format that was visually visible to the human eye, and maintenance staff determined where problems existed, determined and implemented measures to resolve the problems.

In addition, when deciding on countermeasures to solve problems, we use a performance prediction function that predicts performance based on the system model, provide performance data to each individual, and analyze the prediction results to consider improvement measures. It was also done. Analysis using this performance prediction function was performed independently of performance data collection and performance data analysis using the software monitor.

[Point B while the invention is trying to solve the problem]

In the conventional performance improvement method described above, a maintenance person determines the location of a problem based on performance data collected by a software monitor, and determines countermeasures to solve the problem. requires advanced education, and it is difficult to implement appropriate countermeasures. In addition, even if measures are taken to solve performance problems, questions such as ``How will the measures solve the problems?''
Since quantitative predictions cannot be made immediately, in many cases improvement effects have been determined qualitatively and countermeasures have been implemented. For this reason,
If the improvement effect after implementing countermeasures is insufficient, it may be necessary to collect performance data again and repeat the problem, and there were problems in which it took a long time to improve system performance. .

Furthermore, analysis using the performance prediction function was performed independently from performance data collection and performance data analysis using the software monitor, and there was no effective coordination, which increased the burden on maintenance staff. However, it was difficult to implement immediate improvement measures.

The present invention solves these conventional problems, and its purpose is to reduce the burden on maintenance personnel by making it possible to automatically perform the work of considering performance improvement measures to a certain degree. In addition, promptly presenting improvement measures,
The goal is to make implementation possible.

[Means for solving problems]

In order to achieve the above object, the system performance improvement method of the present invention includes a system model used in the performance prediction function in advance in an electronic computer system equipped with a performance prediction function that performs analytical performance prediction based on a system model. system model registration means that inputs the initial system model table and creates an initial system model table; system performance data collection means that collects performance data within the system and stores it in the performance data table; system performance data monitoring means that refers to the performance data table and monitors whether or not a performance problem has occurred in the system; and a system performance data monitoring means that detects a system performance problem by the system performance data monitoring means. system performance predictive control determining means that executes a procedure for determining an improvement measure to improve the detected problem; system performance improvement measure selection means for selecting one of a plurality of improvement measures considered to solve the above problem and notifying the system performance predictive control judgment means; , system model table updating means for updating the initial system model table with the improvement measure selected by the system performance improvement measure selection means and the performance data on the system performance data table to create an updated system model table; , a system performance prediction unit that performs analytical performance prediction based on the updated system model table created by the system model table update unit, and notifies the prediction result to the system performance prediction control determination unit; and the system performance prediction unit. improvement measure display and implementation means for displaying and implementing an improvement measure determined to be optimal by the control determination means; The system performance improvement measure selection means, the system model table update means, and the system performance prediction means are operated repeatedly until there are no improvement measures to be selected by the system performance improvement measure selection means, and performance prediction is repeated to determine performance improvement measures. configured to select.

[Effect]

Improvement measures are registered in advance in the system performance improvement measure selection means for each problematic performance element.

When the system performance data monitoring means that refers to the contents collected by the system performance data collection means and stored in the performance data table detects that a performance problem has occurred in the system, the system performance predictive control judgment means is notified, and the system performance predictive control determination means calls the system performance improvement measure selection means. When the system performance improvement policy selection means is called, it notifies the system performance predictive control determination means of the improvement policy corresponding to the performance element that is the detected problem, and the system performance predictive control determination means responds to the system model The table updating means is called to create an updated system model table based on the selected improvement measure and the performance data on the system performance data table. The system performance prediction means performs analytical performance prediction based on the updated system model table created above, notifies the prediction result to the system performance prediction M'S determination means, and determines whether the notified prediction result is satisfied. In contrast, if the improvement measure is displayed or implemented by the improvement measure display implementation means,
If the prediction result is not satisfactory, the above operation is repeated if there are other improvement measures.

〔Example〕

Next, embodiments of the present invention will be described in detail with reference to the drawings.

Referring to FIG. 1, the embodiment of the present invention includes a system model registration means 3 in which a system model to be used in the performance prediction function 1 is input in advance and an initial system model table 2 is created, and performance data in the system. System performance data collection means 5 that collects and stores in a performance data table 4
and a system performance monitoring means 6 that refers to the performance data table 4 created by the system performance data collection means 5 and monitors whether or not a performance problem has occurred in the system. When a system performance problem is detected by the means 6, a system performance predictive control determining means 7 executes a procedure for determining an improvement measure to improve the detected problem; A system performance improvement measure that is called and selects one of a plurality of improvement measures that are considered to solve the performance problem detected by the system performance monitoring means 6, and notifies the system performance prediction 1IllII determination means 7. Called by the selection means 8 and the system performance predictive control determination means 7, the system performance improvement measure selection means 8
a system model table updating means 10 for updating the initial system model table 2 with the improvement measures selected by and the performance data on the system performance data table 4 to create an updated system model table 9; a system performance prediction means 11 that performs analytical performance prediction based on the updated system model table 9 created by the means 10 and notifies the system performance prediction control judgment means 7 of the prediction result; and a system performance prediction control judgment means 7. and an improvement measure display/implementation means 12 for displaying and implementing improvement measures determined to be optimal.

The system model registration means 3 registers, in the initial system model table 2, a system model required by a performance prediction function 1 that performs analytical performance prediction using a system model.

The system performance data collection means 5 collects various performance data statuses within the summer machine system, such as online response status, bunch zip processing status, central processing unit usage status, main storage processing status, auxiliary storage device usage status, Various queue status.

The usage status of various tables in the system is sampled at predetermined time intervals, and if necessary, it is converted into a predetermined format such as resource usage rate, processing amount per unit time, -processing time per processing unit, etc. Store in performance data table 4.

The system performance data monitoring means 6 refers to the performance data stored in the performance data table 4 at regular time intervals,
Compare it with predetermined performance standards and monitor whether the performance standards are met.

Here, if it is determined that the performance standard is not satisfied, the system performance predictive control determining means 7 is notified together with the problematic performance element that does not satisfy the performance standard.

Next, the operation of the system performance predictive control determining means 7 will be explained with reference to FIG. First, in the problem performance element storage step 71, the problem performance element reported from the system performance data monitoring means 6 is stored in the problem performance element storage area 77, which is an internal table.Next, in the improvement measure retrieval means calling step 72, the problem performance element reported from the system performance data monitoring means 6 is stored in the problem performance element storage area 77, which is an internal table. A zero-order improvement measure selection end determining step of calling the system performance improvement measure selection means 8 using the performance element as a parameter to obtain one of the performance improvement measures and storing it in the improvement measure storage area 78, which is an internal table. At step 73, it is determined whether or not the performance improvement policy obtained from the system performance improvement policy selection means 8 has been selected. If the selection has been completed, the process branches to termination processing step 76. If obtained, the performance prediction calling step 74 updates the system model table updating means 10. The system performance prediction means 11 is executed to obtain the predicted performance when the improvement measures are implemented, and to store it in the predicted performance storage area 79 which is an internal table. Then, in a performance prediction result determination step 75, it is determined according to a certain standard whether the system performance after implementing the performance improvement measures obtained in the performance prediction calling step 74 is satisfactory in terms of performance. If the predicted performance is satisfactory here, the system performance predictive control determination process is ended. If not satisfied, call the improvement measure selection means again Step 7
Return to 2. In the end processing step 76, the problem performance element storage area 77 and the improvement measure storage area 7, which are internal tables, are stored.
8. Problem performance elements stored in each predicted performance storage area 79.

The improvement measure and predicted performance are notified to the improvement measure display implementation means 12.

Next, the operation of the system performance improvement measure selection means 8 will be explained.

As described above, the system performance improvement policy selection means 8 is called by the system performance predictive control determination means 7 with the problem performance element as a parameter, and selects the performance improvement policy that is considered to solve the problem performance element in the performance improvement strategy table, which is an internal table. The improvement measures selected from 80 and notified to the caller can be selected in order from the measures that are most likely to be effective in dealing with the problem performance element by devising the performance improvement measures table 80. . At this time, the selected measures are sequentially stored in another internal table 81, and the same improvement measures are not selected during the - times of processing by the system performance predictive control determining means 7. Also,
When all performance improvement measures corresponding to the performance problem element have been selected, the caller is notified of the selection completion.

FIG. 3 shows an example of the contents of the performance improvement strategy table 80 used by the system performance improvement strategy selection means 8. As shown in FIG. A plurality of strategies are stored that have the potential to improve performance.

Next, system model table updating means 10. System performance prediction means 11. The operation of the improvement measure display implementation means 12 will be explained.

The system model table updating means 10 uses the system performance data table 4 created by the system performance data collection means 5 and the system performance improvement measure selection means 8
The initial system model table 2 is updated based on the performance improvement measures selected, and an updated system model table 9 is created as a system model after implementing the performance improvement measures.

The system performance prediction means 11 calls the performance prediction function 1 with the updated system model table 9 updated by the system model table update means IO as input,
Performance prediction is performed using the system model after implementing the improvement measures, and the system prediction control determining means 7 is notified of the performance prediction result.

The improvement measure display implementation means 12 displays the following information, if the measure determined by the system performance predictive control determination means 7 as a satisfactory performance improvement measure can be implemented dynamically during system operation.
In addition to displaying the information on an external display device such as a system console, implement measures.

In the case of measures that cannot be implemented dynamically, such as changing the physical file arrangement, only displaying on an external display device is possible. In addition, if no improvement measure that can improve the performance has been found, a message to that effect is displayed.

〔Effect of the invention〕

As explained above, the present invention provides performance improvement work that requires advanced technology in the performance improvement work of a computer system, in which performance data collected by a system performance data collection means is analyzed and system performance improvement work based on this data is performed. It is now possible to automatically consider improvement measures to a certain degree, and furthermore, through effective coordination, quantitative estimates of improvement effects, which previously had to be carried out independently, can be calculated incidentally. There is an effect. Therefore, the burden on maintenance personnel is reduced, and improvement measures can be promptly presented and implemented.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the configuration of an embodiment of the present invention, FIG. 2 is a flowchart of a processing example of the system performance predictive control determination means 7, and FIG. 3 is a performance improvement used in the system performance improvement measure selection means 8. 8 is a diagram illustrating an example of the contents of a policy table 80. FIG. In the figure, 1... Performance prediction function, 2... Initial system model table, 3... System model registration means, 4... Performance data table, 5... System performance data collection means, 6... - System performance data monitoring means, 7... System performance predictive control determination means, 8... System performance improvement measure selection means, 9... Updated system model table, 10... System model table updating means, 11 ...System performance prediction means, 12...
Improvement measure display implementation means, 77...Problem performance element storage area, 78...Improvement measure storage area, 79...Predicted performance storage area, 80...Performance improvement measure table.

Claims (1)

[Claims] In a computer system equipped with a performance prediction function that performs analytical performance prediction based on a system model, a system model used in the performance prediction function is input in advance to create an initial system model table. system model registration means for storing the system model; system performance data collection means for collecting performance data within the system and storing it in a performance data table; system performance data monitoring means for monitoring whether or not a performance problem has occurred in the system performance data monitoring means; and when a system performance problem is detected by the system performance data monitoring means, the A system performance predictive control determining means for executing a procedure for determining an improvement measure; and a plurality of system performance predictive control determining means called by the system performance predictive control determining means and considered to solve the performance problem detected by the system performance data monitoring means. system performance improvement measure selection means for selecting one of the improvement measures and notifying the system performance predictive control determination means; and system model table updating means for updating the initial system model table with the performance data on the system performance data table to create an updated system model table; system performance prediction means for performing analytical performance prediction based on an updated system model table and notifying the prediction result to the system performance prediction control determination means; and an improvement measure determined to be optimal by the system performance prediction control determination means. until a prediction determined to be satisfactory in terms of performance is obtained by the system performance predictive control determination means, or an improvement strategy selected by the system performance improvement strategy selection means. System performance improvement characterized in that the system performance improvement measure selection means, the system model table update means, and the system performance prediction means are repeatedly operated until the performance prediction is repeated until the performance improvement measure is selected. method.