Firewall Wizards mailing list archives

RE: Rationale for BSD (I)PF rule order?


From: Smith Gary-GSMITH1 <Gary.R.Smith () motorola com>
Date: Thu, 8 May 2003 14:59:01 -0500

Hi all,

On on hand we have IPF that uses a "last match wins" and on the other
Netfilter/IPtables (Linux) that uses first match wins. The proponents of
first match wins say that it is more efficient, because not every rule has
to be processed. 

Ultimately, the generation of the rulesets has to be done by a human and
humans have to read the rulesets to figure out why there are (or are not)
working. Consider this analogy of Tome Yates in "Building Linux and OpenBSD
Firewalls": Say we want to selectively allow some pets into our apartment,
instead of selectively allow certain packets into our firewall. If we want
to prohibit most pets, but we find that goldfish and iguanas are bearable.
It would be more understandable to say "not pets allowed, except for
goldfish and iguanas" than to say "Goldfish and iguanas are allowed in my
apartment. No other pets are allowed". Eventhough the latter would sound
more natural to a computer, it is human beings who will maintain the pet
rules (or in this case, firewall rules). 

Even with last match counts scheme of IP Filter, you can still force to stop
rule processing at a certain matching rule by specifying the "quick" key
word in the rule. 

Cheers,

Gary


-----Original Message-----
From: Barney Wolff [mailto:barney () databus com]
Sent: Thursday, May 08, 2003 12:38 PM
To: Volker Tanger
Cc: firewall-wizards () honor icsalabs com
Subject: Re: [fw-wiz] Rationale for BSD (I)PF rule order?


On Thu, May 08, 2003 at 02:59:39PM +0200, Volker Tanger wrote:

I was not able to find a rationale for the BSD type of packet filter
application. Where most FW/ACL implementations follow "first match", BSD
usually takes "last match" (if you don't use the "quick" method).

Is there a reason why that was decided this way? Especially as I
currently cannot see advantages for this behaviour, only performance
disadvantages. Can someone enlighten me here?

I can't supply a rationale for last-match, but note that ipfw is first
match, not last.

-- 
Barney Wolff         http://www.databus.com/bwresume.pdf
I'm available by contract or FT, in the NYC metro area or via the 'Net.
_______________________________________________
firewall-wizards mailing list
firewall-wizards () honor icsalabs com
http://honor.icsalabs.com/mailman/listinfo/firewall-wizards
_______________________________________________
firewall-wizards mailing list
firewall-wizards () honor icsalabs com
http://honor.icsalabs.com/mailman/listinfo/firewall-wizards


Current thread: