[e2e] TCP ex Machina
hari at csail.mit.edu
Mon Sep 9 11:20:39 PDT 2013
On Sep 9, 2013, at 12:57 PM, Detlef Bosau wrote:
> Am 09.09.2013 18:15, schrieb Hari Balakrishnan:
>> A RemyCC is described by a set of "match-action" rules mapping an
>> input signal to an action. Each rule has a signal and an action. The
>> signal we explored in the paper has three components: <ack_ewma,
>> send_ewma, and rtt_ratio>, where ack_ewma is the EWMA (low-pass
>> filter) of the interarrival time between new ACKs, send_ewma is the
>> EWMA of the TCP sender timestamps (echoed in the received ACKs), and
>> rtt_ratio is the ratio of the most recent RTT to the smallest seen so
> So, my understanding is correct that the only congestion signalling in
> your algorithm is achieved by
> - the interarrival time of ACKs,
> - the RTT
> - the minimum RTT?
Yes. But please note that the ACKs also tell us the sender's transmission timings of the corresponding packets (segments, if used in TCP). We tried some others but these worked the best. Explaining why that is trickier, and is a question we are investigating currently.
>> The action has three components: <m, b, r>, where new_cwnd = old_cwnd
>> * m + b, and packets are sent (paced) at a rate not exceeding 1
> How do you obtain <m,b,r>?
That's the "superrational" learning stage, described in Section 4.3 of the paper. May I please request you to read that part -- it's actually only less than a page long? Happy to answer more specific questions about the mechanics of the learning. And if you are so inclined, the source code to play with the learning method is available from our web site -- mit.edu/remy
(Please note: the learning stage isn't just doing a simulation of a specific network that is identical to the one under test/evaluation, but is attempting a clever way to efficiently navigate a large design space, while computing an explicit objective function that quantifies the goodness of any given match-action rule.
Thanks for your interest and questions,
> Detlef Bosau
> Galileistraße 30
> 70565 Stuttgart Tel.: +49 711 5208031
> mobile: +49 172 6819937
> skype: detlef.bosau
> ICQ: 566129673
> detlef.bosau at web.de http://www.detlef-bosau.de
More information about the end2end-interest