Paul D. Robertson wrote:

> On Tue, 19 Feb 2002, Martin Peikert wrote:
>>It's not the OS that will solve your problems. The security of an OS is 
>>dependend of the ability and knowledge of it's administrator. If you are 
> 
> It's also dependent on its codebase (size, complexity, design, 
> implementation.)


Right, I forgot to mention that - but I thought it's obvious...

 
>>more familiar with Linux, stay with Linux. If you are interested in a OS 
>>that is focussed on security, try OpenBSD.
> 
> Familiarity is something that should be balanced against homogenity of the 
> environment.  If your security infrastrcture is the same as your server 
> infrastructure, then there's the potential that a single problem will more 
> likely affect both systems.


ACK - it would be better to run _two_ firewalls with _different_ 
operating systems.

> The OBSD work really has more relevence in servers than firewalls, as most 


I cannot agree to that.


> of the exploited services shouldn't be running on a firewall in the first 

> place.  Other than the ICMP kernel bug recently, there's not much that 
> should have affected a well-configured Linux firewall in the last couple 
> years.


Oops - what about the ip_conntrack_ftp bug (see 
http://www.securityfocus.com/archive/1/177070 for more information)

or the bug in Linux 2.4 / iptables MAC match module (see 

http://www.securityfocus.com/archive/1/219180 for more information)?


None of them hare related in any way to the configuration of the system.


Martin

P.S.: It would be sufficient to answer to the list.
-- 
dipl.math. martin peikert                                   Discon GmbH
IT-Security Engineer                                 Wrangelstrasse 100
http://www.discon.de/                             10997 Berlin, Germany

_______________________________________________
Firewalls mailing list
[EMAIL PROTECTED]
http://lists.gnac.net/mailman/listinfo/firewalls

Reply via email to