Empirical Evaluation of Traffic Shaping Algorithms for Time Sensitive Networking

Abstract

Standard Ethernet networks cannot provide solutions to handle latency sensitive applications efficiently. The packet scheduling algorithms like First In First Out (FIFO), Class Based Queueing (CBQ), and others do not provide efficient solutions to Quality of Service (QoS) parameters like end-toend delay, packet loss, and jitter. Time Sensitive Networking (TSN) can be used as a solution to provide QoS to time sensitive applications. TSN has emerged as a future of realtime communication. The main advantage of TSN is that it enables determinism by supporting time critical traffic while the best effort traffic is also present in the network. This paper explores two of the most popular and widespread traffic shaping mechanisms in TSN: Time Aware Shaper (TAS) and Credit Based Shaper (CBS). IEEE 802.1Qbv is used for delivering time assurance using TAS. CBS is a key traffic shaping algorithm to provide bandwidth assurance to the time critical and real time traffic, such as the audio traffic. This paper evaluates TAS and CBS using TSN enabled Network Interface Cards (NIC) with time synchronization, real time kernel and real traffic, which includes time sensitive traffic and elastic background traffic.