Microservice compositions based on the choreography of BPMN fragments: facing evolution issues

COMPUTING(2022)

引用 2|浏览2
暂无评分
摘要
Business Processes (BPs) are commonly used by organizations to describe their goals. However, the existent decentralization found in many organizations forces them to build such BPs by coordinating distributed and fragmented BPs. Within this context, microservices arise as a very interesting and convenient way to address the implementation of such processes due to their low coupling characteristic. In this case, the coordination of such fragmented BPs is usually achieved by means of event-based choreographies. One of the main challenges to be faced by choreographies is their evolution due to the complexity that introduces the need of integrating changes among autonomous and independent partners. We face the challenge of evolving a microservice composition that is globally defined in a BPMN model but executed through a choreography of BPMN fragments. We introduce a protocol to manage the propagation of a change done by one microservice to be integrated into both the BPMN fragments of the rest of the microservices and the global BPMN model. This protocol also supports the negotiation among participants and the automatic suggestion of model adaptations to maintain the functional integrity of the composition. These suggestions are supported by a catalogue of adaptation rules that precisely characterize every possible change and propose actions to be considered by the affected microservices. All the evolution process is done at the modelling level, without managing hard-coded implementations. We have developed specific tools to facilitate the practical adoption of this protocol, and we have validated our work in an experiment with users. We can conclude that the proposed approach is effective to evolve microservice compositions implemented as event-based choreography of BPMN fragments from the local perspective of one partner.
更多
查看译文
关键词
Microservices, Composition, Evolution, Protocol, Bottom-up
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要