Draft paper on connection splitting and BBR (to appear at USENIX ATC)

[Keith Winstein]

Hello BBR team and all,

Our students Gina Yuan and Thea Rossman have written a paper that will appear at the upcoming (and final) USENIX Annual Technical Conference that includes some surprising-to-me results comparing BBRv1, BBRv2, and BBRv3. We have about three weeks before finalizing the camera-ready version of this paper, and we'd love to get any feedback from this group before it's locked. Here is the abstract and current draft for your consideration. Hope this is interesting to you, and thank you in advance for any comments.

Sincerely,

Keith

Abstract: In the 1990s, many networks deployed performance-enhancing proxies (PEPs) that transparently split TCP connections to aid performance, especially over lossy, long-delay paths. Two recent developments have cast doubts on their relevance: the BBR congestion-control algorithm, which de-emphasizes loss as a congestion signal, and the QUIC transport protocol, which prevents transparent connection-splitting yet empirically matches or exceeds TCP’s performance in wide deployment, using the same congestion control. In light of this, are PEPs obsolete? This paper presents a range of emulation measurements indicating: “probably not.” While BBR’s original 2016 version didn’t benefit markedly from connection-splitting, more recent versions of BBR do and, in some cases, even more so than earlier “loss-based” congestion-control algorithms. We also find that QUIC implementations of the “same” congestion-control algorithms vary dramatically and further differ from those of Linux TCP—frustrating head-to-head comparisons. Notwithstanding their controversial nature, our results suggest that PEPs remain relevant to Internet performance for the foreseeable future.