Service configuration automatic optimization method based on space search

Abstract

The invention relates to a service configuration automatic optimization method based on space search. The method comprises the following steps: firstly, acquiring candidate configuration items by using a gridding sampling method under a given configuration constraint condition; then, running a test on the target application, collecting and configuring a performance result corresponding to the target application, and converting the performance result into scalar performance measurement through a utility function; and finally, searching by using a recursive bounded search method to obtain the configuration with the optimal performance. Therefore, under the condition that resources are limited, automatic configuration adjustment and optimization are achieved for wide cloud applications and services.

Description

technical field [0001] The invention relates to an automatic optimization method for service configuration based on space search, which belongs to the field of software technology. Background technique [0002] The number and scale of cloud services are constantly increasing, and the types are diversified, and users need to face more and more configuration parameters. The large number of configuration parameters leads to increasing configuration complexity, which puts great pressure on users, developers and administrators, and is more prone to configuration errors, resulting in performance degradation. Therefore, in order to tap the performance potential of the service and ensure the quality of service, users need to find a reasonable configuration through configuration optimization. However, adjusting the general service configuration becomes difficult due to the following issues. First of all, cloud services are diverse, including data analysis services, database service...

AI Technical Summary

Problems solved by technology

The literature (DanaVan Aken, et al. 2017. Automatic Database Management System Tuning Through Large-scale Machine Learning. In Proceedings of the 2017 ACM International Conference on Management of Data. ACM, 1009–1024.) considers only database management systems and response latency performance goals , but in a cloud computing environment, the configuration tuning process must also take into account various combinations of various services, performance goals, and loads of the application

Second, cloud services are complex

Third, provisioning cloud services has high operational overhead

Literature (Chunqiang Tang, et al. 2015. Holistic configuration management at Facebook. In Proceedings of the 25th Symposium on Operating Systems Principles. ACM, 328–343.) proposed a specific automatic configuration optimization method for distributed application systems, but for various Such cloud services are not universal, and it is difficult to solve the above challenges in the cloud computing environment

In particular, statistical or machine learning models have been proposed for distributed systems, but these models are usually studied based on simulation environments, not suitable for complex cloud computing environments with high-dimensional parameters, and rarely consider the computational overhead of real environments

Images