BBR-S: A Low-Latency BBR Modification for Fast-Varying Connections

Abstract

The new possibilities offered by 5G and beyond networks have led to a change in the focus of congestion control from capacity maximization for web browsing and file transfer to latency-sensitive interactive and real-time services, and consequently to a renaissance of research on the subject, whose most well-known result is Google's Bottleneck Bandwidth and Round-trip propagation time (BBR) algorithm. BBR's promise is to operate at the optimal working point of a connection, with the minimum Round Trip Time (RTT) and full capacity utilization, striking the balance between resource use efficiency and latency performance. However, while it provides significant performance improvements over legacy mechanisms such as Cubic, it can significantly overestimate the capacity of fast-varying mobile connections, leading to unreliable service and large potential swings in the RTT. Our BBR-S algorithm replaces the max filter that causes this overestimation issue with an Adaptive Tobit Kalman Filter (ATKF), an innovation on the Kalman filter that can deal with unknown noise statistics and saturated measurements, achieving a 40% reduction in the average RTT over BBR, which increases to 60% when considering worst-case latency, while maintaining over 95% of the throughput in 4G and 5G networks.

RAPID: A RAN-aware performance enhancing proxy for high throughput low delay flows in MEC-enabled cellular networks

“…On the other hand, as the endto-end RTT increases, it takes more RTTs to reach the BDP of the link, which allows BBR to get a better estimate of the bottleneck bandwidth (i.e., before Cubic introduces large buffering) but causes Cubic to leave the exponential increase phase prematurely. In fact, Cubic's Hystart exits the slow start phase after a certain increase in RTT [33], meanwhile BBR's startup phase continues until it introduces more than two BDPs of inflight data [15]. Due to RTT increase, if BBR does not observe a lower minimum RTT (RTTmin) during a 10 seconds window, it enters a phase called probeRTT.…”

“…Furthermore, in mobile networks, due to the use of deep buffers in base stations, recent CCAs such as BBR that attempt to minimize delay increase are outperformed in terms of goodput by loss-based TCP variants when they compete for bandwidth. Other studies show that, in addition to being RTT unfair, BBR also suffers from bandwidth overestimation issues in fast-varying networks [15] and severe throughput collapse due to high delay variations over millimeter wave (mmWave) wireless links [16]. These issues along with the ever-growing number of CCAs that could potentially compete for bandwidth make it very challenging to solve the self-inflicted bufferbloat issue with a new CCA.…”

Next-generation network transmission control protocol based on deep learning

“…As a result, high-speed connections in Vegas appeared to outperform earlier technologies' sluggish response times. When the transmission speed is decreasing, or when CWND is little, Westwood resembles New Reno [23]. To enable appropriate reactions to diverse events, Westwood's numerous routing strategies vary when CWND rises above a speci c threshold.…”

Rapid: A Ran-Aware Performance Enhancing Proxy for High Throughput Low Delay Flows in Mec Networks

“…Furthermore, in mobile networks, due to the use of deep buffers in base stations, recent CCAs such as BBR that attempt to minimize delay increase are outperformed in terms of goodput by loss-based TCP variants when they compete for bandwidth. Other studies show that, in addition to being RTT unfair, BBR also suffers from bandwidth overestimation issues in fast-varying networks [15] and severe throughput collapse due to high delay variations over millimeter wave (mmWave) wireless links [16]. These issues along with the ever-growing number of CCAs that could potentially compete for bandwidth make it very challenging to solve the self-inflicted bufferbloat issue with a new CCA.…”