From: nick at virus-l.demon.co.uk (Nick FitzGerald)
Subject: DCOM RPC exploit  (dcom.c)

Paul Schmehl <pauls@...allas.edu> replied to Ron DuFresne wrote:

<<snip "corporate networks should have no exposure>>
> Are you really serious?  Recall Slammer?  There were networks that were
> locked down pretty tight.  Slammer couldn't get in, right?  Then one
> developer who got his unpatched copy of SQL inside the network, by
> logging in through VPN with his infected laptop, took the entire network
> down.

Or the remote office which was "too small to be worth getting a 
firewall for", or the contractor or consultant whose laptops are not 
required to meet the same standards as in-house machines brought it in 
from home/another site/etc, and so on and so forth...

> You can't get in to our network on those ports either - unless you're
> already in.  But I can guarantee you that we'll be chasing infected
> boxes down for days after the worm hits.  And we've already patched
> everything that we could patch.  ...

Yep...

> ...  I scan for Slammer every week, because
> every week someone new decides to install SQL unpatched or some stupid
> app that has an unpatched copy of MSDE.  Now I'll be chasing the RPC
> worm around too.

It's kinda sad that "Trustworthy Computing" turned out to mean "keep 
pouring the same bug-ridden shite, pre-any-service-pack-or-hotfix-or- 
any-other-nod-to-security-awareness "redistributables" into everything 
that needs them and caveat emptor...

> You can't firewall 135 inside your network or you'd have no network.

8-)

But that's not _quite_ the issue here.

The issue is that DCOM RPC is silently enabled by default, a very small 
proportion of Windows users use DCOM, about one in a million of _them_ 
actually "need" it to be available "remotely", an even smaller 
proportion of those with it enabled know it even firking exists and 
could be a vector for someone to crap in their shoes, and yet Windows 
Server 2003 _still_ shipped with it installing _enabled and bound to 
all TCP/IP interfaces by default_.

Let me reiterate, in case anyone missed it already -- no-one at 
Microsoft with any clout and any balls understands security.

> The only reason I read lists like this is because I need to know before
> it hits what the next stupid exploit is that I have to deal with.  And
> every one is a royal PITA.  ...

If you want to save yourself much possible pain from this and/or any 
other future MS RPC-related nonsense, please read the excellent "how 
not to write a trustworthy OS vis network services likely to be 
massively unnecessarily exposed to the public sewer known as the 
Internet" paper:

   http://www.hsc.fr/ressources/breves/min_srv_res_win.en.html.en

Actually, it goes by the much more prosaic title "Minimizing Windows 
network services" and has been updated in recent months to include some 
useful, advanced, highly esoteric,  TCP/IP service binding information. 
Why MS makes it so hard to disable, or even better, to selectively 
configure, the binding of services to TCP/IP (in particular) network 
interfaces is beyond me, especially for services that no-one in their 
right mind would knowingly expose to the Internet (oh, of course, I 
forgot -- MS employs the best programmers so they could _never_ make a 
mistake and thereby expose your machine to any unwanted "external" harm 
so we should all enable everything and not worry...)

As this has turned into a rant, I may as well throw this in too...

Is it just me, or is there something sinister, or at least seriously 
disingenuous, in MS arguing on one hand that software developers _need_ 
liability exception under the law "because it is impossible to produce 
error-free software", yet on the other hand never having shown a hint 
of concern that its product's were developed with exceedingly little 
recognition of the fact its own developers may make hugely embarrassing 
errors?  (Of course, the latter is understandable in light of the fact 
that all major legislative bodies have more or less followed the US 
governments cowering to the "but we cannot do it well enough" argument 
and provided the software development industry with its begged for 
liability exclusions...)

OK -- so that was not entirely fair...  Most of the other large 
software developers who also lobby for the "we're too lame to make good 
software so you'd better legally protect our shoddy products" are 
probably similarly disingenuous.

> ...  I put virus and worm writers right there in
> the same pile with spammers.  They're all the scum of the earth.  Clear
> examples of the worst of human nature.

Once upon a time I'd have agreed with you, but of late I feel my 
attitudes to these two groups (malware writers and spammers) changing --
I think you should stop being so nice about the spammers...


Regards,

Nick FitzGerald

Powered by blists - more mailing lists