Weight-Polling Based Task Classification Towards Flexible Computing

With the development of cloud computing and IoT( Internet of Things), we almost arrive "Internet of everything". The number of devices, which are deployed on the network edges, increasing sharply. It resulted in numerous data in sensing layer, and leads to network heavily loaded & high delay. Edge computing has its advantage "close to data source", which can reduce the network latency significantly, but the ability of computing limited. In this paper, we propose a novel weight-polling based task classification scheme towards flexible computing. According to the demand characteristics of resources, tasks are divided into three types: computational, communication and storage. The goal of scheme is to trade off the costs between data transmission of cloud computing and compute of edge devices. The experimental results show that flexible computing can effectively reduce the network delay and response time, improve the resource utilization and task throughput, so as to take into account both the advantages of cloud computing and edge computing, to achieve the fairness of task scheduling.