Vulnerability Development mailing list archives

Re: TCP/IP ISN Prediction Susceptibility


From: Vitaly Osipov <vosipov () WOLFEGROUP COM>
Date: Tue, 13 Mar 2001 10:08:25 -0000

without any further information it looks like reinventing the wheel - e.g.
it is well known that Win NT sequence numbers are easily predictable... Or
just have a look at "nmap -O" output - it gives you a degree of
predictability of those numbers. So only if they have discovered some really
serious flaw in an algorithm of FreeBSD for example (which is considered
"truly random" by nmap :) ).... otherwise it's the same type of media hoax
as a recent "report" about Russian hackers taking over US ecommerce sites
(which in fact is just a restatement of some year old and half-year old
microsoft IIS exploits ;) )

regards,
W.


----- Original Message -----
From: "Solar, Eclipse" <solareclipse () PHREEDOM ORG>
To: <VULN-DEV () SECURITYFOCUS COM>
Sent: Monday, March 12, 2001 9:53 PM
Subject: [VULN-DEV] TCP/IP ISN Prediction Susceptibility


Quoted from http://www.guardent.net/pr2001-03-12-ips.html

Waltham, MA -- March 12, 2001 -- Guardent, Inc., the leading
provider of security and privacy programs for Global 2000
organizations, today released new information regarding a
significant weakness in many implementations of the
Transmission Control Protocol (TCP) that affects a large
population of Internet and network-connected devices.

Tim Newsham, a senior research scientist at Guardent,
discovered a method by which malicious users can close
down or "hijack" TCP-based sessions on the Internet or
on corporate networks. The research, titled "ISN Prediction
Susceptibility", exposes a weakness in the generation of
TCP Initial Sequence Numbers, which are used to maintain
session information between network devices.

Prior to Guardent's discovery, it was believed that TCP
sessions were sufficiently protected from attacks by the
random generation of initial sequence numbers. It is now
known that these numbers are guessable on many platforms,
with a high degree of accuracy. The ability to accurately
guess sequence numbers, combined with readily available
session information, allows for a variety of sophisticated
attacks on computer networks.

It seems that Guardent claims that the pseudo-random ISN
generation algorithm implemented in most TCP/IP stacks
is flawed. Does anybody have more information about this?

Solar Eclipse


Current thread: