Am 11.07.2007 um 20:53 schrieb Bill Marquette:
I know of no official audit of our code. Nor have I ever seen a post
to bugtraq, full-disclosure, or anything on secunia. But take that
for what it's worth...nothing.
A code audit of the GUI/back-end would be pretty nice.
But even if the code was audited, only a specific version (at a
specific point in time) would have that certification.
For such a fast evolving product like pfSense, that would be about as
useful as EAL4-certifying a vanilla-linux-kernel.
I think that once you disallow administration from WAN, have a
restrictive real-DMZ setup that minimizes direct connections even
from the LAN to the firewall (via proxies) and disable DHCP and DNS-
forwarding (and most everything else that is useful in small LANs),
you should be pretty safe.
What's left are vulnerabilities in the FreeBSD-kernel that might lead
to DoS-attacks - or worse. These attacks would have to be carried out
via pure TCP/IP. It's not impossible (didn't OpenBSD have such a
"remote hole" recently?) - but also not very likely to happen very
often.
The PIX is not much different (OK, in theory only) in this respect -
if the FreeBSD core team or Cisco knew of a remote hole in their
respective software, they'd fix it (one hopes).
It's far more likely that a human misconfiguration occurs than a bug
in any of the two systems causes a security-issue.
cheers,
Rainer
--
Rainer Duffner
CISSP, LPI, MCSE
[EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
