lkml.org 
[lkml]   [1998]   [Apr]   [21]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: T/TCP: Syn and RST Cookies
Date
From

Kent Brockman <heathclf@skynet.csn.ul.ie> writes:
>Where could I find out more information about the transactions being
>played twice?

For the record on this one, the researcher in question is
named Mark Smith, and he was, at least a year ago when I last
had contact with him, at MIT working with Nancy Lynch.
His home page, which contains pointers to some of his
work in this area can be found at <http://theory.lcs.mit.edu/~mass/>.

Quick summary for the impatient: T/TCP can get data corruption
and replay conditions under certain types of crash conditions
(A crash condition means a part of the network hickups. It might
be a machine at either end of the link, or it could be a router.)
Even stronger, unless all network transactions obey some
timing assumptions, no protocol that attempts to do what T/TCP does
can work. The open questions as I understand them at this point
are: (1) do the given assumptions get violated in realtity often
enough that we care? (for example, if the probability is signficicantly
lower than the probability of a data corruption that gets past the
checksum screen, then we probably don't), and (2) does T/TCP
as currently specified even work under the assumption that the
timing assumptions in question are obeyed.

Personally, I think T/TCP is a dead issue until these questions are
addressed in a complete and serious manner.

--
Eric Schenk www: http://www.loonie.net/~eschenk
email: eschenk@loonie.net, eschenk@rogers.wave.ca

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu

\
 
 \ /
  Last update: 2005-03-22 13:42    [W:0.032 / U:0.080 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site