Research On Distributed Computing Resource Selection In Wireless Cloud Computing System

The wireless cloud computing system studied in this paper(also known as cloudlet,fog computing,or mobile edge computing in the literature)exists in the vicinity of mobile terminals,which do not need to go through the Internet.Compared with the remote cloud computing system,this system can receive fast realtime interactive response,with lower delay and higher rate.When the terminal enters the system,the system will make full use of computing resources and reduce the delay of computing requests.This paper studies the selection method of computing resources in wireless cloud computing system,and the following results are achieved.Firstly,we design a table-driven computing resource selection method.In the table-driven method,an information table is designed for the computing unit,which stores the information of the idle computing unit in the system.After receiving the computing request from the terminal,the computing unit will query the information table and select the free unit in the system to process the computing request from the terminal according to the information in the information table.Through the simulation and the analysis and comparison of the simulation results,we conclude that the table-driven computing resource selection method reduces the time delay of the mobile terminal in the system and the system performance is good.Secondly,a source-driven computing resource selection method is designed.In the source-driven method,the computing unit looks for idle computing unit in the system by sending packets.The computing unit in the system receives the calculation request.If it cannot process the request in time,it will send the detection packet.The idle computing unit in the system receives the detection packet and sends the feedback packet.After receiving the feedback packet,the computing unit sending the detection packet will send the calculation request to the idle computing unit according to the information carried by the feedback packet.The method is simulated and the simulation results are compared.Thirdly,mobile terminals send computing requests to the system,which have different requirements on time delay.In order to meet the requirements of different types of computing requests on time delay,this paper designs a method of computing resource selection method based on the difference of time delay requirements.Computing request priority is set according to the time delay requirement of computing request,and computing resources in the system are divided.After the computing unit receives the computing request,computing requests of different priorities are sent to the computing resources of this priority for processing.We have carried on the programming simulation to the method,the analysis simulation result,discovered that the request which has the high request to the time delay reduces obviously in the system the time delay.Fourthly,it is unreasonable to mix the computing requests with different computing amounts for small computing amounts.In this paper,a method of computing resource selection based on computation difference is studied.The simulation results show that the time delay of the small computation separately processed is less than that of the small computation in the mixed processing,but the time delay of the large computation will increase.However,compared with the task processing time of the large computation,the increased time delay is acceptable.