----------
Section 1: Very fluffy.
The entire first paragraph struck me as unnecessary,
as well as the prefix to the second. ("...single minded focus...") It
almost looked like one of my intros. :>
The link between the 4 billion processors and internet computing struck me
as being tenuous at best.
"...at a time when end hosts were powerful and routers had limited
capabilities,"
Beginning to sound like a Dairy Queen commerical. And the link between
this and what is essentially a restatement of the 4 billion
microprocessors is again tenuous. Might strike this entire paragraph.
Might want to prefix the list of questions we intend to answer with a
statement like "our goal is to answer the following questions." Having
the questions bluntly tossed in seemed rough.
Picture: "Norther California" [sic]
----------
Section 2: Much better.
Poor Routing Metrics:
"extremely poorly" might be better as just "poorly"
You seem to dislike using contractions; felt that one or two
sentences might flow better with their use.
2nd column, page 3, begins "A number of factors..."
On the topic of conservatism of results,
"Third, our sample hosts are not routers and hence any packet
traversing the path ABC would undoubtedly traverse B's access links twice;
once from ..."
Wouldn't this make our results *less* conservative, given that we are
overstating the loss rate by measuring it twice on the same links?
-------------
Section 3:
A lot of time is spent explaining TCP. Since I can't remember where this
paper is going, this may or may not be a good idea. But it struck me as
potentially odd.
"...timeout which is armed"
1. To be pedantic, you're misusing "which."
2. "Armed?" I seem to remember you laughing about this one.
Usually when I find myself thinking something like this clever, I end up
throwing it out (no matter how hard I laugh at it). This strikes me as a
similar situation.
3.2, 3.3
You use the phrase "the intuition here" in both sections. It's a
very casual phrase to begin with; might consider a variation in one of the
paragraphs.
You don't seem to say what you used as a simulation to gain those
simulation graphs.
------------
Section 4:
Just noticed that the concept of packet rewriting is going to screw with a
lot of people's concepts of security. Example: The CNN daily polls,
which save your IP to keep you from voting twice. (They're not using
cookies).
Sacred Cows: "... require abandoning the view that the network is a black
box." I keep hearing how touchy the E2E people are. This was fine with
me, but struck me as a potential speed bump for dealing with them.
------------
Section 5:
Hmmm... Right after the challenge quote, I've got something to the effect
of "...God bless America!" This might be an aftershock paragraph
from section 1.
----
Side notes:
I always find it interesting how often people cite their own papers. Is
this due to the small number of people in the field, or due to the
familiarity the authors hold with their own work?
------
Overall: Excellent, given the time frame.