TCP RST attack

[Jacques A. Vidrine]

On Tue, Apr 20, 2004 at 01:32:40PM -0700, Dragos Ruiu wrote:
> That's what this thing boils down to imho - the
> space you have to blast through, the time you have to do it in, and 
> the bandwidth/rate available to do it. And there are competing factors,
> and questions about what are the real world values. I'm still waiting
> on final answers...

Consider that on a T1, you can generate 1536 Mbps = ~4800 RSTs per
second.  If you know ((src addr, src port), (dst addr, dst port)),
and assume a 32K window, then you need to send at most about 2^17
RST packets to hit your target.  2^17 / 4800 =~ 27 seconds.

If you have to guess the source port, then we're talking about 2^16
times as many packets needed, which is still `only' about 20 days.  Of
course, the window is sliding during that time... I'm not sure right now
if that makes your chances better or worse :-)

Cheers,
-- 
Jacques Vidrine / nectar at celabo.org / jvidrine at verio.net / nectar at freebsd.org