I’ve been talking for some
time now about the ICS security community’s awareness that control system
vulnerabilities are not enough alone to affect a successful attack on a control
system. There must be some understanding of the process which will be
controlled by the targeted system. Those discussions have centered on the need
for a team of people to conduct a successful attack. The assumption is that the
team would need at least a control system expert and a process expert that
understands the process involved.
My agreement with that basic premise was shaken a bit here
just recently. A reader contacted me with a request for a favor. He worked for
a control system security research firm that was pitching some sort of ICS
security project to a customer. The customer was a chemical processing facility
and the security folks wanted to show a reasonable scenario where a control
system vulnerability could be turned into a method of attack.
Now these ICS folks understood the control system side of
things real well, but they didn’t know enough about the facility to describe a
specific process vulnerability to exploit with their control system knowledge.
So, what were they going to do? Go out and hire a process engineer with
specific knowledge of that type of facility just to make a pitch? Probably not
a good business decision, especially if they couldn’t sell the idea to the
customer.
No, what they did was contact their local friendly chemical
safety/security gadfly who wrote blog posts about chemical safety and control
system security; yep, yours truly. They asked me if I could come up with a
process vulnerability that could be exploited by a cyber attack. They didn’t
want a detailed plan, just a basic scenario to turn loose their ICS experts on
for the purposes of the pitch.
Well, I almost turned them down. The type facility (their
OPSEC was good they didn’t tell me the actual facility) was one that I had
never been in before (there are plenty of those) and I wasn’t really familiar
with the chemistry involved (and there’s lots of those too), so I didn’t have
any immediate ideas about what they wanted.
But, I really do hate to admit I don’t know something about
everything (I try, I really do), so I decided to do a real quick internet
search and see what I could come up with. And sure enough in about 5 minutes I
had a copy of a hazard identification document that was part of a PHA for just
that kind of facility; actually I had PHA documents for three different
facilities in two countries.
So I selected the best file, the one with facility schematic
and P&IDs for the critical processes and in about 10 minutes I found an
interesting process vulnerability. It wouldn’t cause any catastrophic damage or
earth shaking releases, but it would certainly shut the facility down while
they repaired the damage and rethought their process design. Oh, and the
environmentalists would have a PR field day.
So I reported this back to my reader who was very happy
(maybe happy enough to make a small contribution to the blog). And then I
stopped and thought about what we just did and how easy it really was to do.
If this had been a real group intending to attack the
facility, this would be just the first step. But conceptually this is the step
that we in the industry have been saying is the hard part of a cyber attack on
a chemical facility; understanding the process well enough to determine how to
properly attack it. Now there are still lots of things to do to turn this into
a successful attack, but it wouldn’t take much more process knowledge than I
just provided after less than an hour’s research. The rest of it is pretty much
just control system hacking (engineers hate it when I say stuff like that).
Now I am a bit of a process geek and I do seem to have a
knack for finding problem areas in processes (much to the chagrin of some of my
previous bosses), but there are plenty of people out there with chemical
process knowledge and all sorts of radical groups of a wide variety of
persuasions have demonstrated how easy it is to recruit disaffected technical
professionals.
Oh, and one final thing… Why in the HELL would someone put
their PHA documents on the internet? A PHA is nothing more than a detailed look
at the hazards associated with your process and a description of how you plan
to mitigate those hazards. It is essentially an intelligence work up in
preparation for an attack on your facility. Please lock those puppies up tight.
No comments:
Post a Comment