[2474] in SIPB-AFS-requests
Re: SunOS kernel modifications for resisting SYN attacks
daemon@ATHENA.MIT.EDU (Greg Hudson)
Wed Sep 25 03:59:18 1996
Date: Wed, 25 Sep 1996 03:59:00 -0400
From: Greg Hudson <ghudson@MIT.EDU>
To: mhpower@MIT.EDU
Cc: usenet@MIT.EDU, rtfm-maintainers@MIT.EDU, anxiety-maintainers@MIT.EDU,
sipb-afsreq@MIT.EDU
In-Reply-To: "[37560] in Usenet_Meeting"
> Someone claims to have patches to the SunOS kernel that help defend
> against SYN attacks, at the URL
> http://www.netaxs.com/~freedman/syn/.
These look like they might be worthwhile if we were actually under
attack, but I have my doubts about them for normal use. Decreasing
the syn timeout to 15 seconds wouldn't hurt that much, but it's
conceivable that it might cause connections to fail unnecessarily on a
lossy link. As for the second (more important) hack, I have serious
doubts about how well they were tested; it doesn't boost my confidence
that they said "For some reason, the sun4m kernels don't like" what
they did on the sun4c kernels.