Compliant multimedia storage and data extraction from the untrusted and privacy-sensitive edge

Remote surveillance has been presented by gov-erning bodies as a means to control sustainable harvesting of natural resources. Traditional CCTV-based surveillance systems mounted on board fishing boats are currently being deployed in larger scale to address concerns about sustainable use of ocean resources. With this solution arises challenges related to securing forensic evidence, preserving privacy, and commu-nicating multimedia data in the face of low bandwidth and fleeting satellite connections. In practice, it is challenging, if not impossible, to use state-of-art surveillance technologies for efficient and accurate monitoring and control of the thousand-fold commercial fishing boats involved in commercial fishery. In this paper, we outline the need for a domain-specific distributed multimedia system that can handle extraction and storage of data in untrusted and privacy-sensitive environments. We outline the various measures that were considered when implementing a system for storage and delivery of edge data in the fishery industry. We present our work towards a middleware system that ensures secure storage and delivery of multimedia data for forensic purposes, while preserving privacy of people captured on video, and adapting to the difficult conditions that may arise. We conjecture that deploying a domain-specific monitoring system on the edge solution is a feasible alternative to privacy-infringing CCTV surveillance, and show that deployment of our system is possible despite the constrained resources in edge environments.