标题： 面向拓扑的边缘网络微服务架构：部署优化与实现 显示英文标题

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

摘要： 作为一种普遍的部署范式，将微服务架构（MSA）集成到边缘网络中有望提高服务的灵活性和可扩展性。然而，这也带来了由于节点位置分散和复杂的网络拓扑而产生的重大挑战。在本文中，我们提出了一种拓扑感知的MSA，其特点是包含服务、微服务和边缘节点层的三层网络流量模型。该模型详细描述了边缘网络拓扑与微服务之间的复杂依赖关系，将微服务部署映射到链路流量以准确估计通信延迟。在此模型的基础上，我们提出了一个考虑不同类型服务的加权总通信延迟优化问题。然后，提出了一种新的拓扑感知且个体自适应的微服务部署（TAIA-MD）方案，以高效地解决该问题，该方案能够准确感知网络拓扑，并在遗传算法中引入个体自适应机制，以加速收敛并避免局部最优。大量仿真结果显示，与现有的部署方案相比，TAIA-MD将通信延迟性能提高了约30%至60%，并有效提升了整体网络性能。此外，我们在实际的微服务物理平台上实现了TAIA-MD方案。实验结果表明，TAIA-MD在应对链路故障和网络波动方面表现出优越的鲁棒性。 显示英文摘要

摘要： 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.