WebApp Sec mailing list archives

Re: [WEB SECURITY] Re: Oracle in war of words with security researcher


From: Paul Schmehl <pauls () utdallas edu>
Date: Fri, 27 Jan 2006 23:05:42 -0600

--On January 27, 2006 3:59:15 PM -0800 Valkyrie <valkyrie () hacktek com> wrote:

Again, my assertion goes back to failure to have received a logical
response to
the question, "How long is too long to fix your stuff?"

I'll answer it.

1) From the time a vulnerability is made public, a vendor should make a public announcement and provide a workaround or temporary "fix" within 24 hours. 2) From the time a vendor is made aware of a vulnerability, a vendor should make a public announcement that includes workarounds and mitigating factors within 24 hours. 3) From the time a vendor is made aware of a vulnerability until a patch is provided for the *current* (or affected version - if still supported) should never be more than one month. If a vendor can't provide a patch within one month of becoming aware of the vulnerability, they should either hire temporary additional staff or cease development and dedicate staff to the problem so that they can. If fixing a buffer overflow "breaks" your software, maybe you need to go back to the drawing board and learn how to code to begin with.

If more vendors used these benchmarks, developers would learn much more quickly how to spot security problems in their code and avoid the same mistakes in their future work. We're still seeing buffer overflows routinely, for crying out loud. Surely by now *every* programmer is aware of bounds checking??? Surely every programmer knows by now that you don't accept untrusted input without defining the parameters within which it must fall before accepting it???

*Nothing* is more important than providing patches for existing product that is in production and in use by paying customers. Every time a vendor takes longer or sidesteps the issue or doesn't communicate willingly and openly about the problem, they lose credibility. Many times in the past, customers have felt trapped by monopoly software vendors who had no competition and therefore didn't worry about fixing problems. I won't even use or recommend for purchase any security product that has ever had a security flaw in it that wasn't fixed quickly and openly, and if a vendor suffers several such failures, I won't purchase their products *ever*. *Any* of their products. If they don't recognize security problems in their *own* code, how in God's name could I trust them to recognize security problems in *other* vendors' code?

The two most braindead statements of the recent past are:

Microsoft's claim that they had "eliminated buffer overflows in Windows XP" (at the official launch in New York), only to have eEye announce the UPnP buffer overflow one month later

Larry's Ellison's claim that Oracle customers could "keep their Microsoft Outlook, and we will make it unbreakable; and unbreakable means you can't break it, and you can't break in." only to have David Litchfield demonstrate, the very same day, how to break in to 9i and obtain the keys to the kingdom - the administrator account.

Vendors should be paying researchers bounties for doing their work for them. Large bounties.

Apparently only mega-lawsuits will get vendors to change their ways. With all the new laws requiring *customers* to maintain security or pay hefty fines (as was announced today), those lawsuits will soon be forthcoming. Will vendors then *finally* wake up?

Paul Schmehl (pauls () utdallas edu)
Adjunct Information Security Officer
University of Texas at Dallas
AVIEN Founding Member
http://www.utdallas.edu/

-------------------------------------------------------------------------
This List Sponsored by: Watchfire

Watchfire's AppScan is the industry's first and leading web application security testing suite, and the only solution to provide comprehensive remediation tasks at every level of the application. See for yourself. Download AppScan 6.0 today.

https://www.watchfire.com/securearea/appscansix.aspx?id=701300000003Ssh
--------------------------------------------------------------------------


Current thread: