A Dynamic Deep-learning-based Virtual Edge Node Placement Scheme for Edge Cloud Systems in Mobile Environment

Abstract : Edge node placement is a key topic to edge cloud systems for that it affects their service performances significantly. Previous solutions based on the existing information are not suitable for the mobile environment due to the mobility and random Internet access of end users. In this paper, we propose a dynamic virtual edge node placement scheme, in which the edge node placement strategy is generated based on the prediction information. Our placement scheme applies the pay-as-you-go and Spot Instance model of cloud computing, which may allocate the service resources with low cost conveniently and flexibly. What’s more Long Short-Term Memory (LSTM) is implemented to predict the information of end users’ requests and the resources’ prices, endowing the generated placement strategy with the adaptability to the change of end users. At last, a set of hierarchical-clustering-based placement algorithms are proposed, which not only locate virtual edge nodes and allocate their corresponding service resources actively, but also guarantee the service quality of end users with low time complexity. The simulation with trace data shows that compared with K-means-clustering-based placement schemes, our virtual edge node placement scheme can provide users with high-quality service in terms of network delay with relatively low placement cost time-efficiently.

 EXISTING SYSTEM :

 The performance of mobile computing would be significantly improved by leveraging cloud computing and mi-grating mobile workloads for remote execution at the cloud. In this paper, to efficiently handle the peak load and satisfy the requirements of remote program execution, we propose to deploy cloud servers at the network edge and design the edge cloud as a tree hierarchy of geo-distributed servers, so as to efficiently utilize the cloud resources to serve the peak loads from mobile users.

 DISADVANTAGE :

 When there is only one server at each tier of the edge cloud, we formulate the workload placement problem as a mixed nonlinear integer programming (MNIP) problem, with integer variables indicating the locations of workloads being placed and non-integer variables indicating the amount of capacity allocated to execute each workload.

 PROPOSED SYSTEM :

 The hierarchical architecture of edge cloud enables aggregation of the peak loads across different tiers of cloud servers to maximize the amount of mobile workloads being served. To ensure efficient utilization of cloud resources, we further propose a workload placement algorithm that decides which edge cloud server’s mobile programs are placed on and how much computational capacity is provisioned to execute each program. The performance of our proposed hierarchical edge cloud architecture on serving mobile workloads is evaluated by formal analysis, small-scale system experimentation, and large-scale trace-based simulations.

 ADVANTAGE :

 Voice control, recognition assistance and mobile gaming. Many mobile applications are delay-sensitive and require immediate response. Offloading mobile workloads to remote data centers or computing clusters, however, incurs long network transmission latency. Which seriously impairs the mobile application performance