A Fault-Tolerant Deployment Approach for Microservices in the Design Phase of Software Development Life Cycle

Abstract

The concept of microservices architecture has gained significant interest in recent years, primarily due to its capacity to develop large-scale and extensive software systems while offering improved scalability, flexibility, and maintainability. Ensuring fault tolerance during the design phase of the microservices architecture is essential to achieve an efficient and reliable product. This paper introduces a methodology to create a fault-tolerant deployment model of microservices on limited capacitated nodes (servers) in a model-driven architecture. Our approach aims to minimize the total communication cost while maintaining a reliable and fault-tolerant system in the design phase of the software development lifecycle. To evaluate this approach, we conducted various experiments for different capacities of nodes and different numbers of microservice instances. As a result of the experiments, the proposed fault-tolerant algorithm showed improvement rates between 3% to 29% in communication cost compared to the Best Fit, Next Fit, Hungarian and Round Robin algorithms in many cases.