CoEdge: Cooperative DNN Inference with Adaptive Workload Partitioning over Heterogeneous Edge Devices

TL;DR

CoEdge is a system that enables cooperative DNN inference across heterogeneous edge devices by adaptively partitioning workloads, resulting in significant energy savings while maintaining low latency.

Contribution

It introduces a dynamic workload partitioning mechanism for cooperative DNN inference on heterogeneous edge devices, improving energy efficiency and performance.

Findings

Achieves up to 66.9% energy reduction for CNN models.

Maintains low inference latency comparable to existing methods.

Demonstrates effectiveness on a realistic prototype.

Abstract

Recent advances in artificial intelligence have driven increasing intelligent applications at the network edge, such as smart home, smart factory, and smart city. To deploy computationally intensive Deep Neural Networks (DNNs) on resource-constrained edge devices, traditional approaches have relied on either offloading workload to the remote cloud or optimizing computation at the end device locally. However, the cloud-assisted approaches suffer from the unreliable and delay-significant wide-area network, and the local computing approaches are limited by the constrained computing capability. Towards high-performance edge intelligence, the cooperative execution mechanism offers a new paradigm, which has attracted growing research interest recently. In this paper, we propose CoEdge, a distributed DNN computing system that orchestrates cooperative DNN inference over heterogeneous edge devices. CoEdge utilizes available computation and communication resources at the edge and dynamically partitions the DNN inference workload adaptive to devices' computing capabilities and network conditions. Experimental evaluations based on a realistic prototype show that CoEdge outperforms status-quo approaches in saving energy with close inference latency, achieving up to 25.5%~66.9% energy reduction for four widely-adopted CNN models.