Context-aware Computing Offloading Mechanism For Mobile Edge Computing

With the rapid development of Internet of Things and Mobile Computing,mobile edge devices will generate a lot of data.Mobile edge devices are not only data consumers but also data producers.In this scenario,traditional cloud computing is no longer a wise choice since it has some obvious drawbacks.Firstly,the bandwidth of the network would be cloud computing s bottleneck,it is not acceptable for real time computation task on mobile edge devices.Secondly,user privacy is another problem of cloud computing.Mobile Edge Computing is new technology which provides an IT service environment and cloud computing capabilities at the edge of the mobile network,within the Radio Access Network and in close proximity to mobile subscribers.The aim is to reduce latency,to ensure highly efficient network operation and service delivery,and to offer an improved user experience.In mobile edge computing scenario,network latency can be reduced by enabling compu-tation and storage capacity at the edge network,mobile edge devices can perform computation offloading for computing intensive applications to leverage the context-aware mobile edge com-puting service by using real time radio access network information.Since offloading introduces additional communication overhead,a key technical challenge is how to balance between com-putation cost and communication cost to support applications with enhanced user experience,such as lower latency and energy consumption.In this paper,we design a Mobile Edge Computing System composed of Mobile Edge Devices,Monitor Server and Edge Cloud.All input data is from Mobile Edge Devices.Edge Cloud provides computing services for Mobile Edge Devices.Monitor Server is responsible for service information maintenance and monitoring service status.The system architecture is the foundation of computing offloading mechanism.To protect user privacy,Data preprocessing is proposed which includes data clean and data segmentation.Assuming that horizontal segmentation of input data will not affect computing result and the execution time of computing is proportional to the input data size,input data can be divided into multiple data blocks and data block can be divided into multiple data slices.Only one data slice can be executed at one time.In our research,we focus on reducing the energy consumption of mobile edge device and the response time of computation task.And we propose the energy consumption model and response time model.In order to implement dynamic computing offloading,we proposed a dynamic service per-ception method.There are two parts of dynamic service perception method:service filtering and dynamic service updating.Service filtering mechanism enables computing service discovery for computing offloading mechanism.In order to ensure real-time update of service information,this paper puts forward three methods:spontaneous service updating,delayed service updating based on accessing,as well as delayed service updating periodically.The two major purposes of our research are to reduce energy consumption of mobile edge devices and response time of computation task.To reduce the total energy consumption of mo-bile edge device,an energy consumption priority offloading(ECPO)mechanism is put forward.The response time priority offloading(RTPO)mechanism is raised in order to reduce the re-sponse time of one computation task.Combining ECPO mechanism and RTPO mechanism,we propose a dynamic computing offloading mechanism.In mobile edge computing offloading problem,considering the change of network environment,the power of mobile edge device and other factors,making a dynamic decision will work better than making an unchanging decision at the beginning.Finally,we constructs the prototype of context-aware mobile edge computing system,and conducts simulation experiments in normal network scenarios,congested network scenarios,low battery scenarios and limit response time scenarios respectively.The results of simulation indicate that the proposed context-aware computing offloading mechanism based on mobile edge computing can dynamically adjust its offloading strategy based on the energy consumption of mobile edge devices,response time of computing task and the load of edge cloud.