Page 3, second to last paragraph (right after the bullets). The
"lower bound on routing inefficiency" seems a bit too convoluted, and
it's not really clear which bound is the "upper" and "lower" bound on
a negative. My inclination would be to change this to "... study to
estimate the routing inefficiency of the Internet" (probably need to
say internet here, to shift from abstract networks to the actual
beast, anyhow), and let the paragraph later on about how we think we
underestimate the problems stand alone. The argument about which way
our measurements bound things is probably too complex for a single
sentence. An alternative would be to call the study a "conservative
study of routing inefficiency," because that should make it more clear
that we expect things are actually worse than the study will show.
Page 5, Timeouts and fast retransmit bullet. The last sentence
implies, by the use of "during this latter case," that TCP is able to
distinquish the cases of reordered versus dropped packets when it sees
duplicate acks, and that fast retransmit only happens in the case of
drops. I would start the last sentence with "TCP assumes the missing
packet was dropped, and the next in-sequence packet ..."
The smaller nits I found are: (*'s denote my additions/changes)
Page 1, second paragraph, "heirarchical" -- I before E except after C
Section 2, second sentence. We are talking about an abstract system,
so "there are a number of ways *such a* system can be inefficient"
(not *this* system)
Same page, poor routing metrics bullet, "...generally correspond to
organiztion*al* domains..."
Restrictive Routing Policies bullet, near the end, "Consequently*,*
packets sent from of destined *for* smaller..." The comma is
appropriate and breaks up the flow better. I like the word for rather
than to, it seems to fit nicer, even if it doesn't balance from.
Same section, last paragraph, first sentence. "...less that 275Kbps,
or more than four times..." I would strike the "or."
References, first paper, I'm sure "Team, T. N." is happy to see their
name in print. :) I'm guessing this is a bibtex gotcha.
Andy
-- //------------------------------------------------------------------------- // Andy Collins -- KC6YEY acollins@cs.washington.edu // Graduate Student, University of Washington // Finagle's Fourth Law: // Once a job is fouled up, anything done to improve it only makes it // worse. // (collection last updated by Don Woods) //-------------------------------------------------------------------------