[e2e] TCP ex Machina
hari at csail.mit.edu
Mon Sep 9 09:15:03 PDT 2013
On Sep 9, 2013, at 11:25 AM, Detlef Bosau wrote:
> (my apologies for repost)
> I finally managed to read the paper in more detail.
> What I don't yet understand is the RemyCC algorithm.
> To my understanding, this is an end-to-end algorithm, like in
> traditional tcp, employing a multiplicative decrease / additive increase
> So, the tuple (m,b,r) describes the algorithm?
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 far.
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 packet every r milliseconds (assuming equi-sized packets for simplicity).
When an ACK arrives, the sender computes the signal, matches it to the best rule, and performs the resulting action. The rules partition the space of all possible signals into disjoint subspaces, so any input signal observed matches exactly one rule. This sort of structure is general enough to accommodate a variety of signals and actions, not just the ones we proposed. You can see that the structure of the specific match-action rules we investigated makes the resulting RemyCC not linear in its temporal behavior (like TCP), though it is piecewise linear in time.
> 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