A Novel SDN-enabled Edge Computing Load Balancing Scheme for IoT Video Analytics

Edge computing has been designed to deploy resources in the proximity of IoT devices, which reduces latency and network overhead. Nevertheless, resources on edge servers are limited and must efficiently be managed. In this paper, we propose a novel software-defined networking (SDN)-based scheme to balance the computation resource requests among a network of edge servers aimed at supporting IoT video analytics streaming applications. In the proposed solution, programmable switches periodically report the IoT video streaming workload forwarded to each edge server. This information is then used at the SDN controller to estimate the incoming and outgoing traffic load at edge servers and balance IoT video streaming among them, by updating routing tables at the programmable switches. The performance of the proposed solution is evaluated and compared to related schemes through extensive simulations using the Mininet emulator. Obtained results show that the proposed solution can reduce up to 21 % of average latency with 20 % load saving in each edge server, compared to deterministic and random-based related solutions.