Research On Resource Allocation For Multiple Virtual Machine Online Migration Tasks In Edge Cloud

With the continuous advancement of cloud computing technology,edge cloud technology has been an extension and innovation of the traditional cloud computing model,providing different service applications for more and more users.There are some obvious characteristics for edge cloud data centers,such as small scales,limited scope of service and so on.What's more,edge cloud data centers can receive and response to user requirements in real time.While the traditional cloud data centers are far from users and have wider service range.Every edge cloud data center has limited coverage areas and will not provide service for user who exceed its coverage area.Due to user mobility,there appears such circumstances that users exceed the coverage area of edge cloud data centers leading to the loss of Qo S.Online virtual machine migration among different edge cloud data centers located at different geographic locations is an important strategy for providing efficient service with low latency for mobile users.Because the bandwidth resource is more limited among edge cloud data centers compared to the traditional cloud data centers,the virtual machine migration among edge cloud data centers is more challenging than in cloud data centers.In the actual scenarios,there are usually many concurrent online virtual machine migration tasks carried on while the existing researches only focus on a single virtual machine migration task.On the other hand,in the actual scenarios,the completion of virtual machine migration needs to consume the bandwidth resource among different edge cloud data centers,meanwhile the service on the virtual machines are still provided for users.In such case,there appears the trade-off between migration time and quality of services for users(Qo S).Existing researches haven't considered how to ensure Qo S during the process of the online virtual machine migration.For the above aspects,the main work of this article is as follows:1)Different from previous work,given the bandwidth and deadlines of all migration tasks and other problem parameters,this paper studies the bandwidth resource allocation scheme of multiple concurrent virtual machine migration tasks in order to maximize the Qo Sunder the constraint of migration time.2)The corresponding heuristic optimization algorithm is proposed to solve our research problem.By comparing this algorithm with the benchmark algorithm,the experimental results show that the proposed heuristic resource allocation algorithm is superior to the benchmark algorithm in the overall experimental solution.3)In addition,the online resource allocation scheme is also studied.We proposed an heuristic online resource allocation method and the corresponding benchmark algorithm.Through the simulation experiment analysis,it can be found that the online resource allocation scheme of the heuristic method is better than the benchmark algorithm in the aspects of solution stability and rationality of results.