Real-time deep learning based traffic analytics

The increased interest towards Deep Learning (DL) technologies has led to the development of a new generation of specialized hardware accelerator [8] such as Graphic Processing Unit (GPU) and Tensor Processing Unit (TPU) [1, 2]. Although attractive for implementing real-time analytics based traffic engineering fostering the development of self-driving networks [5], the integration of such components in network routers is not trivial. Indeed, routers typically aim to minimize the overhead of per-packet processing (e.g., Ethernet switching, IP forwarding, telemetry) and design choices (e.g., power, memory consumption) to integrate a new accelerator need to factor in these key requirements. Previous works on DL hardware accelerators have overlooked specific router constraints (e.g., strict latency) and focused instead on cloud deployment [4] and image processing. Likewise, there is limited literature regarding DL application on traffic processing at line-rate [6, 9].