Research And Implementation Of Adaptive Evolution Framework For Microservices Application Driven By Requirements

Distributed software systems in dynamic and open Internet environment are typically decentralized,heterogeneous and evolving.To satisfy the continuous changes of the user requirements,web applications need to understand and manage requirements in a flexible and adaptive way,and then build reusable software units to enhance the evolution of software.In the one hand,web services are widely used in distributed software systems because of the good interoperability,cross-platform and loosely coupled characters.In the other hand,with the advent of microservice architectures,the modularization of distributed systems has been well resolved.So how to construct a service sequence that satisfies users' changing requirements based on the microservice architecture will be of utmost importance.Aiming at the problems above,a requirement-driven microservices application adaptive evolution framework RMAE is proposed.The purpose of this paper is to study the design and implementation of the framework and then construct an adaptive platform for service dynamic evolution based on RMAE.The platform can integrate third-party services on the Internet,and support the evolution of services to meet the diverse requirements of users.The main innovations in this article include:(1)An adaptive evolution framework for microservices application driven by requirements is proposed.By means of the control objectives feedback loop(CO-FL),adaption feedback loop(A-FL)and monitoring feedback loop(M-FL)in DYNAMICO reference model,it maps onto the process of route delegation,microservices scheduling and monitored execution,and thus the evolving requirements from users is satisfied.(2)RDM and RCT models are designed to support the adaptability of the framework.The RDM model describes the structure of the ontologies required by the services and the relationship between the ontologies,while the RCT model describes the implementation of user requirements.(3)A service dynamic evolution strategy is proposed.By matching the best service of each RCT node,the service execution sequence that satisfies the user requirements can be obtained,and the process includes service clustering,service filtering,and service ranking.(4)We built an adaptive platform for service dynamic evolution in open environments and verified its adaptability.Then,the performance and development efficiency of RMAE were evaluated,and we applied it to the actual development cases.