Fuzzy inference for service migration strategy

Abstract

Service migration is an important approach for service availability and system survivability in a security incident. When a system is under attack and some platforms have been compromised, the services executed on those platforms must be migrated to other platforms in order for them to be continuously provided to users. Service migration strategy is the guideline and high-level decision regarding what (e. g., service programs, the service state and the data space) are moved from one platform to another and how. In this paper, we present a fuzzy inference system to determine the most appropriate strategy for service migration in a security incident scenario. Our approach uses expert knowledge as linguistic reasoning rules and takes as input the current system state such as the damage degree of the service programs, the complexity of those service programs, and the available network capability to securely transfer service programs and data to a new platform. Simulations show that the fuzzy inference system is effective in determining the most appropriate strategy for service migration given the current system state and environment information.