A Performance Study of CPRI over Ethernet with IEEE 802.1Qbu and 802.1Qbv Enhancements

Abstract

There has been a debate on whether or not Ethernet, a highly cost effective technology, could meet the stringent latency and jitter requirements imposed by CPRI. To facilitate the discussion, we conducted simulations to understand how Ethernet performs when carrying CPRI traffic with the two Ethernet enhancements, currently being standardized by IEEE, namely frame preemption (802.1Qbu) and scheduled traffic (802.1Qbv). Our simulation results led to two conclusions: 1) Ethernet networks with or without frame preemption, regardless of being shared or dedicated to CPRI traffic, can not meet the CPRI jitter requirement of 8.138 ns, confirming a widely hold belief; 2) Ethernet with the enhancement of scheduled traffic in conjunction with a well defined scheduling algorithm could significantly lower or even completely remove jitter thus could meet CPRI jitter requirement. To the best of our knowledge, this paper is among the first to study the performance of Ethernet with IEEE 802.1Qbu and 802.1Qbv enhancements and to demonstrate by simulation that Ethernet with scheduled traffic could meet CPRI jitter requirement.

Analysis and modeling of asynchronous traffic shaping in time sensitive networks

“…Some works in the literature elaborate the requirements and implementation of the real-time scheduled traffic in TSN networks. For instance, references [3] and [4] give examples of applying Ethernet and scheduled TSN to in-vehicle and wireless communication systems, emphasizing the importance of scheduling. The evaluation in [3] proves that Ethernet is able to transport the traffics mixing of different vehicle functions but scheduling is necessary in the overload situations.…”

“…The evaluation in [3] proves that Ethernet is able to transport the traffics mixing of different vehicle functions but scheduling is necessary in the overload situations. In [4], results show that it is difficult for conventional Ethernet to fulfill the jitter requirements of Common Public Radio Interface (CPRI), while this problem could be solved by implementing enhanced scheduled traffic. The paper [5] proposed simulations of scheduled traffic in Audio Video Bridging (AVB) network.…”

Modeling Time Aware Shaping in an Ethernet Fronthaul

“…For the latter, minimizing (or completely removing), any latency variation (including contention-induced) is important for achieving proper accuracy. Some work in the literature has been carried out to analyze the effects of contention with TSN features [13][14] in an Ethernet fronthaul and its effects on KPIs. Specifically [13] is an implementation of a time-aware shaper (TAS) in NS3 focusing on CPRI performance, while [14] presented an initial TAS implementation in Opnet with limited features.…”

“…Some work in the literature has been carried out to analyze the effects of contention with TSN features [13][14] in an Ethernet fronthaul and its effects on KPIs. Specifically [13] is an implementation of a time-aware shaper (TAS) in NS3 focusing on CPRI performance, while [14] presented an initial TAS implementation in Opnet with limited features. This work focuses on the scheduling of traffic in an Ethernet fronthaul network and presents a model implementation in Opnet of a TAS based on IEEE 802.1Qbv.…”

Scheduling in an Ethernet fronthaul network

“…A number of papers have analysed the effects of contention, with [12][13][14] and without [15] scheduling, in an Ethernet fronthaul and its effects on KPIs. In this work, timeaware shaper (TAS) and SP models are built and configured, respectively, in OPNET.…”