Research And Implementation Of Mobile Computation Offloading In Cloud-end Mode

In recent years,smart mobile terminals and cloud computing have developed rapidly.On the one hand,people can do more and more operations on mobile phones.On the other hand,“cloud” plays an increasingly important role in people's lives.The development and impact of both have a great deal of overlap and will become closer and closer.The advent of smart phones emergence has led to the rapid development of hardware and software,with smaller devices,better performance and more functions.However,just like the failure of Moore's Law,mobile terminals are constrained by the conditions of equipment.People have been facing various problems with insufficient resources of equipment.For example,when too many applications are running,some Android devices suffer from stuttering due to insufficient memory,or the battery life may deteriorate after the battery ages.With the development of technologies and networks,people are increasingly demanding on computing power,resource concentration,and resource utilization,and cloud computing has emerged.The cloud-end convergence computing mode and the corresponding technology platform provide a completely new solution to the problem of operation experience caused by insufficient resources of mobile terminal devices,this is,by migrating part of computing tasks to servers with more computing resources such as CPU and memory,in order to enable the use of cloud's computing,storage and network resource,lighten computational load on mobile devices,improve application performance and reduce energy consumption of mobile devices.We call it computation offloading.To achieve computation offloading,the crux lies in how to migrate local computing tasks and implement cloud computing agent.The existing academic research is carried out at all levels of the system,but it is difficult to achieve the landing of the technology because of the limitations of their own soft and hardware factors.In this paper,we study the computation offloading technology for Android applications,design and implement computation offloading technology for application developers based on OpenStack cloud platform.Applications on devices are developed with computation offloading SDK,which is use for cost estimation,decision making and computation offloading.OpenStack virtual machines are built based on Android x86 in cloud,which carry the operation of computing agents and the storage of target code.on devices.Based on this,we use the GBRT algorithm to optimize the load balancing of computing offloading requests,and verify the effectiveness of the computing offloading technology in reducing a running app's response time and energy consumption through experiments.