Edge-First Resource Management for Video-Based Applications: A Face Detection Use Case.

The edge computing paradigm introduces a hierarchy of multiple processing elements between the edge devices, the gateways, and the cloud endpoints, in order to address the Internet-of-Things (IoT) challenges in a scalable way. In order to support the computational demands of latency-sensitive video applications and efficiently utilize the available network resources, we present an edge-based resource management methodology for serving video processing applications in an IoT environment. In this letter, we propose a scalable solution for providing low-latency video analytics at the edge level, while maximizing the quality of service under device and network constraints. As use case, we evaluate the proposed methodology on a face detection video processing application. The experimental results show that the proposed methodology satisfies specific time thresholds, comparing to cloud-only, local-only, and other video-optimized offloading techniques, while taking into consideration the video resolution. [ABSTRACT FROM AUTHOR]