Re: [Full-disclosure] Getting Off the Patch

[Michael Krymson <krymson@xxxxxxxxx>]

For once in this thread, thank you for a concrete example! :)

As a follow-up pair of related questions...

1) Do you find it taxing (or costly) to have to always defend those moments
where patching is not used? For instance, any auditor or evaluator of your
security is going to immediately pounce on missing patches.

2) Somewhat similarly, you won't ever "pass" on any automated tools, which
will certainly incur manual labor costs to iron out false positives, right?
I'm not saying we should use and trust such automated (usually
vulnerability) scans, but that's a reality of so many firms. I imagine every
pen test or vuln scan or other audit is pretty intensive and requires lots
of back and forth in an environment that has chosen other controls vs
patching.


The dangerous part of this paradigm about operations vs patching (and the
original article in this thead!) is how truly difficult it is to get other
people on board, from deeply technical to managerial and shallow. Even my
mom would raise an eyebrow to hear that firm XYZ may not necessarily patch.
It might be worthwhile to just patch and continue to do what is easy to
understand and move on with, rather than incorporate so much complexity and
head-spinning control?

One of the very first bullet points on any corporate or home computer
security regiment is to patch. And then we have people saying it's ok not to
patch if/when/blah.

How does that get us any further? How does that make management feel any
better about what we offer?


<- snip ->
*Re: Getting Off the Patch*
------------------------------
*From*: "Cor Rosielle" <cor () outpost24 com>
*Date*: Wed, 19 Jan 2011 12:08:48 +0100
------------------------------

I don't agree with the statement: "From a security standpoint, patching is
better than not patching.  Period.".

Sometimes patching is the right solution, often it is not. Since some asked
experiences from larger companies, here is one:

In 2001 I was responsible for maintaining all kinds of systems and services
at a telephony and internet provider. One morning a list with our company
name in it was mentioned in the radio news bulletins. We also found our name
in this list in articles in newspapers. A "hacker" found we didn't install a
patch on one of our web servers which ran IIS 5.0 on Windows 2000. The patch
was available for about 3 months and the hacker claimed he could attack our
server because of this specific patch missing. Because of his "ethical
standards" he did not try to really attack the server but just publish about
the shameful patching policies in large companies.
Three years later, when a new website was developed, the server was
replaced. The patch was still not installed. Even though we were in the
radio news bulletins and news papers and the world knew about our vulnerable
server, no successful attack occurred in all that time.

O yes, they tried. I've seen it in our log files. But because other defense
mechanisms, the known exploit did not work on our server. And since the web
server was in no danger without this patch, we decided installing the patch
would be a higher risk than leaving the server as is.

I did not know about the OSSTMM in those days. If I did, I could have
explained why patching is not always the best solution: it interferes with
your operations. And if it influences you operations, you better control it.
Not blindly execute it and install the patch using an automated update
process, but actually control it.
So the first thing to do is to decide if applying a patch is useful at all.
And often it is not! E.g. why would you even consider to install MS10-085 on
a http-only web server (MS10-085 apparently fixes an error in TLS
handshake)? (Don't flame me on this one, it's just meant as an example). And
if you concluded a patch is useful, then you decide if you do need to
install it, or if it is not really necessary to install it. And if you
install it, then decide if you do it manually, in a controlled manner, or
use an automated update process, in an uncontrolled manner. The OSSTMM helps
you to realize that an automated update process increases the attack
surface, which better be controlled.
Another option of course is to blindly install it, because you trust your
vendor (you know, the one that provided buggy software in the first place).
Then your not controlling, but trusting. Read chapter 5 in the OSSTMM to
find out if that is wise for this particular vendor, or not.

Bottom line is that patching interferes operations and therefore, from a
security standpoint, it either has to be controlled or trusted. It is not
always true that patching is better than not patching. So I would slightly
like to rephrase that statement:
"From a security standpoint, thinking is better than not thinking.
Period.". (Now would it surprise you if I told you that critical thinking is
one of the starting points of the OSSTMM??.

Cor Rosielle
Chief Technology Officer

_______________________________________________
Full-Disclosure - We believe in it.
Charter: http://lists.grok.org.uk/full-disclosure-charter.html
Hosted and sponsored by Secunia - http://secunia.com/