Empowering Microservices: A Deep Dive into Intelligent Application Component Placement for Optimal Response Time

Abstract

Microservice architecture offers a decentralized structure using componentization of large applications. This approach can be coupled with Edge computing principles: applications with stringent response time can benefit from different deployment options. However, it is crucial to gain profound insights into correlations between the deployment of distributed application components and the response time, especially from an application perspective. For correct placement decisions, it is important to evaluate the impact of small functions’ placement and their interactions across the Edge–Cloud Continuum. This paper investigates the response time from an application perspective, considering the componentization using microservice architecture. Unlike the existing application placement approaches, we present extensive simulation results, illustrating the impact of service chains and marginally considered Application Programming Interface Gateways placement. Numerical evidence depicts that the design and placement of microservice-based applications could counter the common perception that Edge resources are always suitable for user-perceived response time. Further, we also present an experiment involving a componentized application and its optimized deployment in an actual testbed. Our findings and design guidelines inform effective component placement decisions while considering infrastructure constraints as well.