BBR-With Enhanced Fairness(BBR-EFRA): A new enhanced RTT fairness for BBR congestion control algorithm
Charles Kihungi Njogu,Wang Yang,Humphrey Waita Njogu,Adrian Bosire
DOI: https://doi.org/10.1016/j.comcom.2022.12.015
IF: 5.047
2022-12-23
Computer Communications
Abstract:Towards the end of 2016, the Google research team proposed and developed a new state-of-the-art TCP congestion control algorithm called Bottleneck Bandwidth and Round-trip propagation time(BBR). When deployed on various Google internal servers, BBR attained higher throughput and low latency performances on modern-day and sophisticated networks than traditional congestion control algorithms like Cubic, Reno, or Vegas. Unlike conventional congestion control algorithms, BBR controls data transmission by maximizing delivery rate and minimizing the round-trip time(RTT), therefore maximizing bandwidth utilization and improving throughput and latency delay performances. However, some experiments have reported persistent queue formation and massive packet retransmissions rate in the bottleneck link, which happens to be the main cause of unfairness between different RTT flows. BBR prefers and favors long RTT over short RTT flows; therefore, long RTT flows are allocated more bandwidth when competing with short RTT flows in the bottleneck link. It was also noted that even the minor difference between the two competing flows could be a source of throughput imbalance and unfairness. The dominance of long RTT flows is the central origin of the high queuing delay, packet loss and retransmission rates, and even severe BBR vulnerability that malicious users can exploit to obtain a larger share of bandwidth simply by increasing the RTT(delay). Therefore, we propose a BBR-With Enhanced Fairness(BBR-EFRA) to mitigate this major concern challenge. Our proposed algorithm adaptively controls the congestion window(CWND) by adjusting the bandwidth delay product(BDP) values of each BBR flow during data transmission based on buffer queue status computation. Our proposed approach guarantees different RTT flows to compete for the available bottleneck bandwidth more equally, hence improving the BBR fairness issue. We evaluated our algorithm on the NS3 simulating environment. BBR-EFRA shows improved RTT fairness by more than 16%, reduced retransmission rate by more than 20%, and queuing delay by more than 18%, and finally increased Jain's fairness index by more than 1.3 times compared with other recently published BBR variants like an adaptive congestion window of BBR(BBR-ACW). Therefore, with BBR-EFRA, short RTT flows can fairly compete for the available bandwidth even with higher RTT differences under various network scenarios.
computer science, information systems,telecommunications,engineering, electrical & electronic