Method and system for evaluating performance of server

Abstract

The invention discloses a method and a system for evaluating the resource utilization rate of a server. The method comprises the following steps of: sampling the resource utilization rate of the server according to set sampling time and sampling frequency; performing statistic analysis on sampled data to obtain 90 percent of peak value, 95 percent of peak value, 99 percent of peak value and the maximum peak value; and according to 90 percent of peak value, 95 percent of peak value, 99 percent of peak value and the maximum peak value, which are obtained by statistic analysis, calculating evaluation indexes of the resource utilization rate of the server by using a weighted mean method. The invention also discloses the system for evaluating the resource utilization rate of the server. The system comprises a resource utilization rate sampling unit, a sampled data statistic analysis unit and a utilization rate evaluation index calculation unit. The method for evaluating the resource utilization rate of the server provided by the invention has the advantages that: by acquiring the resource utilization rates of different periods of the current on-line server, key weighing parameters in an evaluation process are determined, and the unified evaluation of the resource utilization rates of the servers of all application systems is realized.

Description

technical field [0001] The invention relates to the field of performance evaluation, in particular to an evaluation method and system for server resource utilization. Background technique [0002] With the expansion of information system business applications, various application systems have put forward higher requirements for digital centers that manage and store data. However, with the deepening of the complexity and application scope of the application system, the performance of some existing servers in the established data centers may not meet the requirements of the application and urgently need to be transformed. However, the server asset life cycle is not over and cannot be eliminated. If some existing servers are completely replaced and transformed, the investment cost of the transformation will be too high. In addition, in the process of perfecting and developing the data center, some business applications may have high configuration and low utilization rate. [...

AI Technical Summary

Problems solved by technology

Although these standards provide basic methods for evaluating the performance of various servers, the performance evaluation standards for different types of application servers are different, and there is no general performance evaluation method to uniformly evaluate the resource utilization of different servers in various business application systems

Moreover, before using various standards for estimation, it is necessary to set up a strict test environment according to the requirements of each standard, which cannot achieve universality

[0005] In addition, the user's use of the application server may be concentrated in certain time periods, so the server will have unevenly distributed peaks

However, when various standards evaluate the servers of various applications, they do not accurately consider the impact of different input volumes at different time periods during actual operation on server resource utilization.

If the resource utilization evaluation results at the peak of the system are used to deploy future resources, it may cause over-allocation of resources, while using the average evaluation results may lead to insufficient resources at peak usage times, resulting in performance degradation

Images