ELTOS: Energy-Latency Trade-Off Optimization Strategy for Microservice Placement in Edge Environments

Abstract

The deployment of microservices in edge environments is increasingly critical to support latency-sensitive and data-intensive applications such as IoT analytics, real-time monitoring, and smart city services. Edge infrastructures are highly heterogeneous in terms of compute capacity, communication latency, and energy efficiency. Furthermore, the geographical distance between edge nodes may introduce non-negligible energy consumption for data transfers, which is often overlooked in traditional placement strategies. While prior works focus on minimizing response time, there is a need to adapt and extend such strategies for energy efficiency in edge computing. This paper proposes an Energy-Latency Trade-off Optimization Strategy (ELTOS) for microservice placement in edge environments. ELTOS formulates and solves a cost-based optimization for microservice placement on heterogeneous edge servers. The main goal of this optimization algorithm is to minimize the transmitted data size between microservices on different edge servers, saving energy while ensuring a suitable level of end-to-end latency for each request. Experimental evaluations using real edge computing infrastructure demonstrate the efficiency of the proposed ELTOS compared with previous work. A comparison result of energy consumption shows that ELTOS outperforms the previous method in 70% of all pairwise comparisons; therefore, ELTOS consumes less energy while maintaining QoS performance level. ELTOS enhances the median, 95th, and 99th percentile response times on average by 6.03%, 7.47%, and 3.30%, respectively.