Search

In an edge computing environment, in order to provide highly accessible services and computing resources to nearby users, edge servers are usually deployed on base stations and other types of equipment. However, due to the limited storage space of edge servers, it is very difficult to manage services with a failure. Therefore, the rapid restoration of services in the edge computing environment will become an important means to ensure the resilience of the system. Among them, service replacement with similarities within the edge server is one of the effective technologies to ensure system resilience. In this paper, we regard the discovery and scheduling problem of replaceable services replacement as the discovery and scheduling component, and develop an approach based on replaceable service form the app vendor’s perspective for solving the none replaceable services environment. We have evaluated our approach in a real experimental environment. The results show that in the case of large mirroring, the DAS approach can effectively reduce the recovery time required by the system due to failure.