Optimal Placement Of Mobile Edge Computing Servers And Resource Allocation In Edge Networks

Mobile edge computing(MEC),as an extension of the cloud computing paradigm to the edge network,is a promising solution to provide resource-intensive and time critical applications to mobile devices.It overcomes the obstacles of traditional mobile cloud computing by offering ultra-short latency and less core network traffic.In this work,we investigate the challenges of the architecture of MEC to deliver cloud services to the edge,and also propose efficient solutions.Since different placement schemes of MEC servers would produce various network performances,especially for the access delay and network reliability.How to efficiently deploy MEC servers among the heterogeneous infrastructures in the edge network,and on the basis of the optimal placement of MEC servers,how to efficiently deploy Virtual-machine Replica Copies(VRCs)supporting multiple applications among numerous MEC servers in the edge network are significant problems to be investigated.By applying Software-Defined Networking(SDN)techniques to provide flexible and programmable management in the edge network,we propose Enumeration-based Optimal Edge Server Placement Algorithm(EOESPA)and Ranking-based Near-Optimal Edge Server Placement Algorithm(RNOESPA)to obtain the optimal placement of MEC servers in the edge network minimizing the average access delay and ensuring high network reliability at the same time.As corroborated by extensive simulation results,RNOESPA reports the nearly optimal placement of MEC servers with lower computational complexity than EOESPA,even than that of the famous K-medians clustering algorithm(KMCA).What's more,the performance of RNOESPA in both average MEC server access delay and network reliability outperforms KMCA,much closer to the optimal solution.After the optimal placement of MEC servers in the edge network,we also have proposed several optimal service deployment algorithms to minimize the average response time in MEC architecture and the service deployment cost in the edge network.Besides the Optimal Enumeration Service Deployment Algorithm(OESDA)as benchmark,we design Latency Aware Heuristic Service Deployment Algorithm(LAHSDA)with much lower computation complexity than EOESPA.To enhance the performance of LAHSDA on minimizing the average response time,Clustering Enhanced Heuristic Service Deployment Algorithm(CEHSDA)is proposed.We also develop Substitution Enhanced Heuristic Service Deployment Algorithm(SEHSDA)to avoid falling into local optimal solutions.As corroborated by extensive simulation results,the performance of SEHSDA is much closer to that of EOESPA compared with LAHSDA and CEHSDA.Note that CEHSDA also outperforms LAHSDA,and both are better than a general Greedy Service Deployment Algorithm(GSDA).