AN EFFICIENT SERVICE MIGRATION MODEL BASED ON IMPROVED GENETIC ALGORITHM IN MOBILE EDGE COMPUTING ENVIRONMENT

摘要

Mobile edge computing (MEC) reduces network operation and service delivery delay by providing IT service environment and cloud computing capability at the edge of mobile network. However, in mobile edge computing environment, resource constrained mobile servers may not be able to achieve efficient computing requirements, and thus cannot guarantee the quality of service execution. Therefore, it is necessary to select the adjacent MEC servers to get computing power support. This paper focuses on service migration in mobile edge computing environment. Considering factors such as mobile server monitoring cost, service execution cost and data transmission cost, a threshold-based edge server selection strategy is designed and an efficient service migration model related to communication distance, network bandwidth and other factors is constructed. A genetic algorithm incorporating back learning and Levy flight mechanism is adopted to solve the service migration model. Experiment results show that the edge server selection strategy and service migration model proposed in this paper have obvious advantages in convergence speed, convergence accuracy, energy consumption and accuracy.