基于多目标遗传算法的微服务粒度识别方法  

Microservice Granularity Identification Method Based on Multi-Objective Genetic Algorithm

作　　者：袁海铭 YUAN Haiming(Faculty of Information Engineering and Automation,Kunming University of Science and Technology,Kunming 650500,China)

机构地区：[1]昆明理工大学信息工程与自动化学院,云南昆明650500

出　　处：《电视技术》2022年第5期67-72,81,共7页Video Engineering

摘　　要：微服务架构因其高度的自治性与可扩展性被大型企业广泛使用,而微服务架构中微服务粒度问题一直以来都是学术界与工业界亟待解决的问题,目前为止仍然缺乏行之有效的方法来识别微服务粒度。本文基于多目标遗传算法中的NSGAII算法,以敏捷方法中用户故事依赖作为输入,通过对微服务粒度的识别得到一组具有高内聚、低耦合性的微服务,并通过比较NSGAII算法与其他三种多目标遗传算法的性能指标,来对优化效果做出讨论。结果表明,所提出的方法优于传统的微服务粒度识别方法,所得到的结果可以为相关从业人员在对微服务粒度进行评估时提供参考。Microservice architecture is widely used by large enterprises because of its high degree of autonomy and scalability. The issue of micro-service granularity in micro-service architecture has been an urgent problem to be solved by academia and industry. So far,there is still no effective method to identify the granularity of micro-service. In this paper, based on the NSGAII algorithm in the multiobjective genetic algorithm, using user story dependency as input in the agile method, a set of micro-services with high cohesion and low coupling are obtained by identifying the granularity of micro-services. The performance of NSGAII algorithm is compared with other three multi-objective genetic algorithms to discuss its optimization effect. The results show that the proposed method is superior to the traditional micro-service granularity identification method, and the results can provide reference for relevant practitioners in evaluating the micro-service granularity.

关 键 词：微服务架构 粒度 多目标 遗传算法 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论]