Dynamic Least-cost Task Scheduling for Enabling Ubiquitous Sensing Service in Edge Computing

Sensor as a Service(SaaS) is introduced by providing ubiquitous sensing services along with prosperity of Internet of Things (IoT) [1]. It comprises billions of mobile devices equipped with sensors which can sense, communicate and compute. With more capabilities, devices could share some computation tasks which are used to be taken by the cloud servers or the edge servers. To achieve the most benefit of energy consumption and delay, we propose Dynamic Least-cost Task Scheduling(DLCTS) mechanism for enabling on-demand ubiquitous sensing service by leveraging edge computing which brings compute and storage resources on or close to devices. Through virtualization of sensors required in a region, we could adapt task assignment by changing mappings between virtual sensors and devices dynamically to the movement of devices and incidents. Simulation results show that the proposed algorithm achieves great cost performance without paying extra expense of quality and decision-making time. And it proves that the scheme is suitable for densely-distributed devices and large-scaled for multiple sensing tasks.