A Heuristic Services Binding Algorithm to Improve Fault-Tolerance in Microservice based Edge Computing Architecture

Chao Lei, Hongjun Dai

Abstract: In microservice architecture, Mobile Edge Computing (MEC) has been widely developed to improve Quality of Service (QoS). This distributed technology deploys different microservices to intricate network environment. A fault-tolerant deployment and scheduling scheme is significant for overall architecture to enhance robustness. Especially in some time-sensitive service scenarios, any network fluctuation caused services invocation failures may lead to irreparable loss. This work introduces a heuristic based services binding algorithm to improve fault-tolerant MEC in microservice architecture with the help of Cache-enabled Edge Nodes. This challenge is modeled as a constrained optimization problem on services composition and network nodes graphs. In addition, this work proposes a graph based state and action value functions to heuristically generate solutions. In overview system, the suitable combination will balance the performance and robustness.