EdgeBench: A Workflow-Based Benchmark for Edge Computing

Edge computing has been developed to utilize heterogeneous computing resources from different physical locations for privacy, cost, and Quality of Service (QoS) reasons. Edge workloads have the characteristics of data-driven, latency-sensitive, and privacy-critical. As a result, edge systems have been developed to be both heterogeneous and distributed to utilize different computing tiers’ resources and features. The unique characteristics of edge workloads and edge systems have motivated EdgeBench, a workflow-based benchmark aiming to provide the ability to explore the full design space of edge applications and edge systems. EdgeBench is both customizable and representative. It allows users to customize the workflow logic of edge workloads, the data storage backends, and the distribution of the individual workflow function to different computing tiers. To illustrate the usability of EdgeBench, we implement two representative edge workflows, a video analytics workflow, and an IoT hub workflow that represent a large portion of today’s edge applications. Both workflows are evaluated using the workflow-level and system-level metrics reported by EdgeBench. We show that EdgeBench can effectively discover the performance bottlenecks and provide improvement implications for the edge workloads and the edge systems.