Multi-Path Service Function Chaining for Mobile Surveillance of Animal Husbandry

Abstract

Animal husbandry is the pillar industry in some ethnic areas of China. However, the communication/networking infrastructure in these areas is often underdeveloped, thus the difficulty in centralized management, and challenges for the effective monitoring. Considering the dynamics of the field monitoring environment, as well as the diversity and mobility of monitoring targets, traditional WSN (Wireless Sensor Networks) or IoT (Internet of Things) is difficult to meet the surveillance needs. Mobile surveillance that features the collaboration of various functions (camera, sensing, image recognition, etc.) deployed on mobile devices is desirable in a volatile wireless environment. This paper proposes the service function chaining for mobile surveillance of animal husbandry, which orchestrates multi-path multi-function (MPMF) chains to help mobile devices to collaborate in complex surveillance tasks, provide backup chains in case the primary service function chain fails due to mobility, signal strength, obstacle, etc., and make up for the defects of difficult deployment of monitoring facilities in ethnic areas. MPMF algorithm models both mobile devices and various functions deployed on them as abstract graph nodes, so that chains that are required to traverse various functions and hosting mobile devices can be orchestrated in a single graph-based query through modified and adapted Dijkstra-like algorithms, with their cost ordered automatically. Experiment results show that the proposed MPMF algorithm finds multiple least-costly chains that traverse demanded functions in a timely fashion on Raspberry Pi-equipped mobile devices.