Firewalls - Ease of Use and Maintenance?
-Hammer-
bhmccie at gmail.com
Thu Nov 10 15:39:29 UTC 2011
OK. Right off the bat you know I can't and won't. But in some places it
is common practice to make sure agreements are in place to make sure all
parties are protected based on how a product is expected/designed to
perform. I can't say more than that. Realize I'm speaking about things
that are solely on the vendor. Not "Did you configure the ACL properly?"
What you can Google is the names of companies who have settled out of
court against various trolling lawsuits vs the names of companies that
are still in litigation. There is a mix of both manufacturer/vendor and
end customer. It all depends on the case.
This shouldn't surprise you. If Toyota makes a defective brake and you
slam into someone else, your insurance covers you. Eventually, if the
issue scales out to the point that it is obvious that Toyota made a
defective brake and it is not your fault, some insurance companies
collectively will go to the government or directly to the manufacturer
for compensation. This is no different. If you sell me a FW and it
catches on fire thru no fault of my own and then the public finds out
that FWs are catching on fire all over the place, it's a good bet that
that FW vendor will be getting some lawsuits. If a FW vendor reports a
product to work a certain way and instead thru a massive vulnerability
or development oversight it does not the same applies. Software.
Hardware. Physical (fire). Logical (vulnerability). I'm not saying that
it happens all the time and I'm not even saying it's a general practice.
What I'm saying is it happens. And depending on your business vertical
it could be a very real consideration.
COMPLETELY 100% MADE UP HYPOTHETICAL SCENARIO:
I put a FW in. I put proper L3 ACLs in. I block 443 inbound. I didn't
say I block HTTPS. I block 443. I test it by telnetting from the
Internet to 1.1.1.1:443 and I am unable to connect. Looks good. A month
later our CEO is surfing the Internet. Thru a development oversight in
the product, when I NAT or PAT him to the Internet his source port is
not pulled from the Ephemeral range but is instead sourced as port 443.
He of course goes to sites riddled with Malware because that's what CEOs
do. They click on links. So the Malware website initiates a new TCP
session to destination port 443 with his NATted IP. The state table has
an entry for that IP and 443 and even though this is a new TCP session
the FW lets it thru. The malware site bad guys are able to retrieve
confidential information about a merger and publish it. The other
company that we were merging with sues us because the information is
leaked to the public and adversely impacted their stock value.
Everything in the above paragraph is able to be documented thru
forensics and it is indisputable that the FW was properly configured and
should have blocked it but didn't. The FW did NOT perform as
advertised/designed. This is NOT the fault of me or my company. If a few
thousand dollars is at stake nothing may come of this. If tens or
hundreds of millions of dollars are at stake I promise you that our
lawyers will be contacting the manufacturer whose product did not
perform as advertised. They will compensate (in one way or another) us
for our losses. It's a big ugly world full of lots of lawyers.
-Hammer-
"I was a normal American nerd"
-Jack Herer
On 11/10/2011 09:14 AM, Richard Kulawiec wrote:
> On Thu, Nov 10, 2011 at 08:52:22AM -0600, -Hammer- wrote:
>
>> The other high cost of "free" that people sometimes overlook is
>> liability.
>>
> Please point to an instance (case citation, please) where a commercial
> firewall vendor has been successfully litigated against -- that is, held
> responsible by a court of law for a failure of their product to provide
> the functionality that it's claimed to provide.
>
> ---rsk
>
>
>
More information about the NANOG
mailing list