Abstract: Because of the corresponding delay and communication cost associated with the use of cloud service,mobile edge computing(MEC) which is closer to mobile users has become the main technology for processing computing-intensive and delay-sensitive applications.Small data centers located on the edge of network are called “cloudlets”,which can provide computing resource for nearby mobile devices.The service delivery delay can be reduced significantly by using MEC.However,in the edge network environment composed of mobile micro-clouds,load balancing directly affects the response time of tasks.In order to improve the quality of service for users,this paper proposed a task offloading and cooperative load balancing mechanism.The mechanism includes a latency-aware target selection strategy (LATS) for mobile users and a collaborative load-balancing strategy (CLB) for mobile cloudlets.LATS chooses the best task migration object for mobile users according to the current load information of cloudlets.CLB uses balls-in-bins theory and it can balance the task loads with extremely limited information.Simulations and evaluations demonstrate that the proposed mechanism can effectively reduce the system delay and load gaps,as well as the communication and computing cost.

Key words: Balls-into-bins theory, Computing and communication cost, Mobile edge computing, User task offloading, Workload balancing