The threat isn't entirely hype. Stuxnet basically proved that dropping a grid is do-able.
Not sure how you'd do individual companies, but the power grid as a whole is a lot more fragile than you might think.
The industry is taking things seriously, but there are a lot of problems to sort out. Fixing legacy network architectures takes time, and a lot of critical systems never get patched. That's before you factor in espionage or 0day.
Didn't Stuxnet target very specific Siemens SCADA systems, unless our grids all run the same controllers and the same software (which they might well do, I have no idea) - surely an attack like that would be difficult to scale?
Oh no... http://www.energy.siemens.com/hq/en/power-transmission/ http://www.energy.siemens.com/hq/en/fossil-power-generation/ http://www.siemens.co.uk/en/news_press/index/news_archive/20...
What I meant was, how diverse are the control systems for national grids. If they are highly diverse then they are going to be more difficult to target, if they are all the same generation kit from the same vendor then the threat will be higher.
Which one is it?
So your point is valid, but diversity doesn't help if the overall resilience of the grid is so low that attacking a few carefully chosen targets could cause a massive blackout.
Typically you might only need to cause a few targeted failures at the right time, and if you do this in a nasty way you can cause a cascading failure, taking out key transmission lines or other nodes in the network.
* http://en.wikipedia.org/wiki/Cascading_failure
* Structural Vulnerability of the North American Power Grid - http://arxiv.org/pdf/cond-mat/0401084v1.pdf
* Risk-aware Attacks and Catastrophic Cascading Failures in U.S. Power Grid - http://www.ele.uri.edu/nest/paper/Globecom_2011.pdf
It's a very "Dr Evil" topic, I know...
But the idea that someone was actually planning to do this to the Olympics opening ceremony probably was hype.
The source of the info isn't stated, but it's more likely to be a case of "forensics panic" or on-site network sensors going off than anything else.
For Critical National Infrastructure this is handled through an organisation called CPNI[1], that runs information exchanges. Where data is regulated by the government's own security framework this is handled by the Cabinet Office[2] and by a section of GCHQ called CESG[3].
> deploying things like SELinux (one good thing the NSA has done)
Are you suggesting that GCHQ should go around installing SELinux on peoples' computers, or that they should tell others to do it? There's actually a pretty comprehensive (if typically bureaucractic) information assurance framework in use throughout .gov.uk.
> doing free pen-testing
CESG (part of GCHQ) have a full service catalogue and they do provide certification for private sector suppliers through the CESG CHECK Scheme[4].
> advising companies on how to architect critical networks and so on.
Again, that's more Cabinet-officey type stuff, but CESG does provide guidance in the form of Good Practice Guide (GPG) documents, here's[5] an example of one.
GCHQ does actually do a lot of good stuff, but Olympic power switch threat management probably isn't the best example.
[1] - http://www.cpni.gov.uk/
[2] - https://www.gov.uk/government/publications/security-policy-f...
[3] - http://www.cesg.gov.uk/Pages/homepage.aspx
[4] - http://www.cesg.gov.uk/servicecatalogue/CHECK/Pages/index.as...
[5] - https://www.gov.uk/government/uploads/system/uploads/attachm...
We don't, so either: it's not that easy, or no one has worked out how to do it yet.
What it doesn't indicate is that spying on everyone helps at all. What GCHQ should be doing is working with infrastructure providers to make sure all their security software is up to date, deploying things like SELinux (one good thing the NSA has done), doing free pen-testing, advising companies on how to architect critical networks and so on.