does anyone know if the scanning is sequential once a range is chosen or is it random within a range? e.g., 1.1.1.1 1.1.1.2 1.1.1.3 etc or 1.1.1.89 1.1.1.33 1.1.1.12 etc -----Original Message----- From: John Dvorak [mailto:john@dvorak.net] Sent: Monday, August 11, 2003 5:57 PM To: NANOG Subject: Re: RPC errors On Mon, 11 Aug 2003 17:33:33 -0400 Kevin Houle <kjh@cert.org> wrote:

--On Monday, August 11, 2003 02:26:40 PM -0700 Mike Damm <MikeD@irwinresearch.com> wrote:

...

The DCOM exploit that is floating around crashes the Windows RPC service when the attacker closes the connection to your system after a successful attack. Best bet is to assume any occurrence of crashing

...

RPC services to be signs of a compromised system until proven otherwise.

http://www.cert.org/advisories/CA-2003-19.html

That's good advice. Many of the known exploits cause the RPC service to crash after the exploit is successful. I'll point out that not all exploits cause the service failure. So, the absence of an RPC service failure is likewise not an indicator that a vulnerable machine has escaped compromise.

Kevin

Interestingly, we have clear examples of boxes which were not infected but on which RPC services did crash. This may suggest that the worm also takes advantage of the unrelated RPC DOS vulnerability (2000 and XP) which I believe MS has still not patched. John