The Application of ChatGPT for Identification of Microservices

Abstract

Identification and definition of microservices is one of the most important aspects of systems based on the microservices architecture. If the services are designed poorly, the overall performance and efficiency of the system might be even worse than with traditional systems. There are many approaches for proper identification of microservices and their boundaries. However, these approaches are usually carried out by software architects, meaning that the overall success of the system design depends on their skills, abilities and understanding of the entire system. This comprehensive task can be quite demanding, which leaves room for oversights and errors. Luckily, new technologies emerge daily, as well as new uses for those technologies, which make such difficult tasks a little bit easier. In this paper, utilization of the popular ChatGPT large language model for analyzing software requirements and identifying microservices is explored. Three different examples are presented along with recommended solutions, showing that usage of GPT for analyzing software requirement can be useful, but used with caution because of its drawbacks.