[e2e] [tcpm] RTTM + timestamps
rs at netapp.com
Mon Jan 17 14:57:34 PST 2011
again thank you for these references!
I wanted to bring this particular issue up, to discuss if an improved RTO calculation should be brought into a hypothetical I-D talking about improved signaling using existing options.
The one currently used by Linux has the merit of extremely wide deployment, and being conservative (only improves one particular aspect over the current standard calculation).
For non-technical issues, "Eifel" algorithms offer less good choice for any such future I-D work... (restrictions on IPR prevent the use in commercial implementations unfortunately).
However, I think most of these scenarios would also benefit from obtaining good feedback about the RTT samples (with the additional side-info, if the RTT was obtained off a reordered or retransmitted segment). But current signaling doesn't yield that feedback back from the client.
From: SCHARF, Michael [mailto:Michael.Scharf at alcatel-lucent.com]
Sent: Montag, 17. Jänner 2011 21:52
To: Scheffenegger, Richard; tcpm at ietf.org; iccrg at cs.ucl.ac.uk; end2end-interest at postel.org
Cc: Matt Mathis; Mark Allman
Subject: Re: [tcpm] RTTM + timestamps
I happened to look at the RTO calculation and RTO spike issue a long time ago. For whatever it is worth, I scanned that old work and extracted some references that may or may not be well-known. A number of algorithms have indeed been developed to address these issues (e. g., Linux stack), and several papers tried to further optimize the timer calculation in particular for mobile environments, including amongst others:
R. Ludwig und K. Sklower. The Eifel Retransmission Timer. ACM SIGCOMM Computer Communications Review, 30(3), 2000, pp. 17-27 [this is not the actual Eifel algorithm]
H. Ekstroem and R. Ludwig, "The Peak-Hopper: A New End-to-End Retransmission Timer for Reliable Unicast Transport," Proc. IEEE INFOCOM, 2004
K. Jacobsson, H. Hjalmarsson, N. Moeller and K. H. Johansson, "Round-Trip time estimation in communication networks using adaptive Kalman filtering"
A. Kesselman, Y. Mansour, "Optimizing TCP Retransmission Timeout", Springer LNCS 3421, 2005
I. Psaras, V. Tsaoussidis, "Why TCP timers (still) don't work well", Computer Networks, Volume 51, Issue 8, 6 June 2007
And, BTW, a young wannabe-TCP researcher once tried to systematically understand the RTO spikes resulting from RFC 2988:
Michael Scharf, Marc C. Necker, Bernd Gloss: "The Sensitivity of TCP to Sudden Delay Variations in Mobile Networks", Springer LNCS 3042, 2004, pp. 76-87
If a better RTT estimation or RTO calculation was indeed needed, these papers might contain some interesting starting points.
More information about the end2end-interest