Eris: An Online Auction for Scheduling Unbiased Distributed Learning Over Edge Networks

Jinlong Pang, Ziyi Han, Ruiting Zhou, Renli Zhang, John C. S. Lui, Hao Chen

Abstract: The emergence of edge intelligence has made smart IoT services (e.g., video/audio surveillance, autonomous driving and smart city) a reality. To ensure the quality of service, edge service providers train unbiased models of distributed machine learning jobs over the local datasets collected by edge networks, and usually adopt the parameter server (PS) architecture. However, the training of unbiased distributed learning (UDL) depends on geo-distributed data and edge resources, bringing a new challenge for service providers: how to effectively schedule and price UDL jobs such that the long-term system utility (i.e., social welfare) can be maximized. In this paper, we propose an online auction-based scheduling algorithm Eris, which determines the data workload, the number and the placement of concurrent workers and PSs for each arriving UDL job, and dynamically prices limited edge resources based on current resource consumption. Eris applies a primal-dual framework which calls an efficient dual subroutine to schedule UDL jobs, achieving a good competitive ratio and pseudo-polynomial time complexity. To evaluate the effectiveness of Eris, we implement both a testbed and a large-scaled simulator. The results demonstrate that Eris outperforms and achieves up to 44% more social welfare compared to state-of-the-art algorithms in today's cloud system.