Title:Topology-aware Microservice Architecture in Edge Networks: Deployment Optimization and Implementation

Abstract:As a ubiquitous deployment paradigm, integrating microservice architecture (MSA) into edge networks promises to enhance the flexibility and scalability of services. However, it also presents significant challenges stemming from dispersed node locations and intricate network topologies. In this paper, we have proposed a topology-aware MSA characterized by a three-tier network traffic model encompassing the service, microservices, and edge node layers. This model meticulously characterizes the complex dependencies between edge network topologies and microservices, mapping microservice deployment onto link traffic to accurately estimate communication delay. Building upon this model, we have formulated a weighted sum communication delay optimization problem considering different types of services. Then, a novel topology-aware and individual-adaptive microservices deployment (TAIA-MD) scheme is proposed to solve the problem efficiently, which accurately senses the network topology and incorporates an individual-adaptive mechanism in a genetic algorithm to accelerate the convergence and avoid local optima. Extensive simulations show that, compared to the existing deployment schemes, TAIA-MD improves the communication delay performance by approximately 30% to 60% and effectively enhances the overall network performance. Furthermore, we implement the TAIA-MD scheme on a practical microservice physical platform. The experimental results demonstrate that TAIA-MD achieves superior robustness in withstanding link failures and network fluctuations.