[ NNSquad ] Re: [OT?] NN definition(s?)

[Richard Bennett <richard@bennett.com>]

On Sun, 2007-11-11 at 11:23 -0500, Fred Reimer wrote:

> Forging packets and sending RST's is bad.  The use of a new
> technology such as P2P in and of itself is not.  I had to FTP
> Fedora 8 from usf.edu just yesterday because I'm on ComCast and I
> got a whopping 4K download speed with bittorrent.  I've never
> used it for any illegal activity, and actually just installed it
> specifically do download Fedora.  Does this tick me off?  You
> bet.

Point of information, I'm on Comcast with the lowest-tier residential
account and at this moment I'm downloading the Fedora 8 DVD over Azureus
with a reported D/L speed of 500 kilobytes/sec, which is 4
Megabits/second if my arithmetic is correct. This is the advertised
maximum d/l speed for accounts of this type.

The reports on Comcast's traffic shaping by fairly responsible parties
have established that Comcast doesn't interfere with BitTorrent
*downloads*, only with BT seeding. So if you've got slow download
speeds, something else is going on. One common problem is the failure to
map the port your copy of BT uses on each computer in your home network
on your NAT box. 

So I have two questions: 

1) Does anybody else think Comcast interferes with BitTorrent
*downloads*?

and 

2) Has it actually been established that RST spoofing is a clear
violation of NN, in some way that RED isn't or real-time SYN dropping
isn't? 

It seems to me that the purpose of RST spoofing is to bring the user
into line with the TOS, by the indirect means of moving him to the
bottom of the BT tracker's list of seeders by bandwidth. If the TOS are
legitimate, so is the means of bringing about compliance, hence no
violation.

Like I said before, we need more light and less heat.

RB