OCTOPINF: Workload-Aware Inference Serving for Edge Video Analytics

TL;DR

OCTOPINF is a workload-aware inference system for edge video analytics that optimizes resource allocation and scheduling to meet real-time requirements, significantly improving throughput and robustness in dynamic edge environments.

Contribution

We introduce OCTOPINF, a novel inference serving system that employs fine-grained resource management and a spatiotemporal scheduling algorithm for edge video analytics.

Findings

Achieves up to 10x throughput increase over baselines.

Demonstrates improved robustness in dynamic edge scenarios.

Effectively balances workloads between edge devices and servers.

Abstract

Edge Video Analytics (EVA) has gained significant attention as a major application of pervasive computing, enabling real-time visual processing. EVA pipelines, composed of deep neural networks (DNNs), typically demand efficient inference serving under stringent latency requirements, which is challenging due to the dynamic Edge environments (e.g., workload variability and network instability). Moreover, EVA pipelines also face significant resource contention caused by resource (e.g., GPU) constraints at the Edge. In this paper, we introduce OCTOPINF, a novel resource-efficient and workload-aware inference serving system designed for real-time EVA. OCTOPINF tackles the unique challenges of dynamic edge environments through fine-grained resource allocation, adaptive batching, and workload balancing between edge devices and servers. Furthermore, we propose a spatiotemporal scheduling algorithm that optimizes the co-location of inference tasks on GPUs, improving performance and ensuring service-level objectives (SLOs) compliance. Extensive evaluations on a real-world testbed demonstrate the effectiveness of our approach. It achieves an effective throughput increase of up to 10x compared to the baselines and shows better robustness in challenging scenarios. OCTOPINF can be used for any DNN-based EVA inference task with minimal adaptation and is available at https://github.com/tungngreen/PipelineScheduler.