An Integrated Top-down and Bottom-up Task Allocation Approach in Social Sensing based Edge Computing Systems

With the advance of mobile computing, Internet of Things, and 5G networks, social sensing based edge computing (SSEC) systems have emerged as a new computation paradigm where people and their personally owned devices collect and process sensing measurements about the physical world at the edge of networks. In this paper, we focus on the task allocation problem in SSEC where rational edge devices are motivated by incentives to collectively accomplish the computation tasks in the system. Several unique challenges exist to solve this problem: (i) the edge devices often do not share the complete context information (e.g., CPU, memory usage) in the task allocation process due to privacy concerns; (ii) the edge devices are rational actors who may have competing objectives with the application; (iii) the application server and edge devices are usually owned by different entities, making the coordination in task allocation more challenging. This paper develops a novel integrated Top-Down and Bottom-Up (TDBU) task allocation framework to address these challenges. In particular, TDBU incorporates abottom-up game-theoretic model that allows the edge devices to specify their task preferences in a way that maximizes their payoffs. It also incorporates atop-down control model that ensures the performance of the applications using control theory. The TDBU was implemented on a real-world edge computing testbed that consists of heterogeneous devices (Jetson TX1, TK1 boards, Raspberry Pi3). We compared the performance of TDBU with state-of-the-art baselines through a real-world social sensing application. The results showed that our solution significantly outperformed the baselines in various application settings.