I'll try sending to the whole list instead of just Ryan :-(

---------- Forwarded message ----------
From: chris derham <[email protected]>
Date: Fri, Apr 8, 2011 at 1:30 PM
Subject: Re: [Owasp-modsecurity-core-rule-set] A Recommended Base
Configuration - SecRuleEngine
To: Ryan Barnett <[email protected]>


All,

Was on leave so only just joining in the fun discussing the options.

I have only been using mod security for the last 4 months. I can still
remember installing it onto an apache server, and then running it and seeing
nothing. I remember that it seemed very strange that it wasn't turned on.
Comparing with other pieces of software, I do not know of any that you have
to install, and then to actually switch on for them to work.

I don't think that I am about to add anything to the prior discussion, but
having seen all the other posts, which all brought up valid arguments, I
guess I'd agree that off makes no sense. detect only seems like a prudent
default. At least if that is the default, people can monitor the logs and
then switch on when happy of no false positives

One thing I did take issue with in the discussions was the following

> Your argument assumes that enabling the
> SecRuleEngine is **the** ideal scenario, this may not always be the
> case. While it may be a common practice, stopping malicious traffic is
> a sure fire way to advertise your running a WAF, which then leads to
> attackers modifying their payloads to elude your ruleset/engine which
> may or may not be subsequently caught. By only monitoring suspicious
> payloads, malicious users are less likely to send obfuscated attacks
> intended to bypass ModSecurity.

1) Please understand I just want to discuss this point, and am not
bad-mouthing the poster.
2) surely given the potential for problems in websites, e.g. xss that is no
such thing as "only monitoring suspicious payloads" - any part of the
website could be compromised and as such there isn't a safe/unsafe split.
3) if there is a flaw in a webserver, and the bad-guys want to abuse your
server with it, then when the bad-guys spot that it doesn't work on your
server, they will deduce you have a WAF. They are quite bright after all. At
this point, the cat and mouse game begins. As I said, I've only been working
with mod security for a few months, but surely in this case, the only thing
to do is to hope that

   - they get bored and go somewhere else
   - your white list rules stop the bad guys while they are attempting to
   get in
   - your proactive monitoring catches anything that they do manage to get
   through

So I guess what I am trying to ask, is should you only detect parts of the
website with mod security, or protect it all? We just went live with a
system and we used the latter

Thanks

Chris
_______________________________________________
Owasp-modsecurity-core-rule-set mailing list
[email protected]
https://lists.owasp.org/mailman/listinfo/owasp-modsecurity-core-rule-set

Reply via email to