Why Flame Can't Be the wWrld's Most Complex Malware

Run the numbers, folks: Flame is probably just low-hanging fruit

The most remarkable thing about the Flame worm is that it's dominating global news. Not that Flame isn't newsworthy. Most people are still unaware of the scope and scale of online espionage. It's new to the audience, therefore it is news. But why now? Let's put Flame into perspective.

Flame can take over computers, log keystrokes and mouse movements, record screenshots, turn on cameras and microphones without turning on red lights, copy and exfiltrate data to its masters - but so can any number of $200 criminal malware kits.

Flame is modular - but so is any good software.

There's a few new-sounding tricks involving Bluetooth and the like - but again that's surely evolutionary rather than revolutionary.

Until now, the reference point for complex malware - at least in public - has been Stuxnet, the worm that supposedly sabotaged Iran's uranium enrichment program in 2009 and 2010.

Stuxnet probably cost just a few million dollars to produce, according to Patrick Gray, presenter of the Risky Business security podcast.

"Ralph Langner, a prominent Stuxnet researcher, he estimated that it took five to ten developers around six months to create it, so probably less than a few million dollars," he said in a speech to Security Summit 2011 (MP3) in South Africa last year.

Gray had run these estimates past "a few people who would know", and he reckons they said Stuxnet would have been "significantly cheaper".

Flame is 20 times the size of Stuxnet, we're told, or 40 times if all its modules are loaded. But as Sophos senior technology consultant Graham Cluley points out, that's just a byte count - and Flame seems to be coded in a way that would naturally result in bloat.

For the sake of my argument, I'll be conservative and guess that Flame took 10 person-years to write. Call it $5 million, given typical defence-contractor margins.

Mikko Hypponen, F-Secure's chief research officer, told the AusCERT conference earlier this month that there's literally hundreds of US malware-writing jobs being advertised today with US defence contractors, and presumably hundreds in all the other technologically-advanced nations too.

These developers are being hired to join others to do work has been going on for years, representing tens of thousands of person-years and billions of dollars.

Ten person-years suddenly looks... insignificant.

There must be hundreds of worms like Flame - many stockpiled for future use, sure, but many in active service, spying for their masters.

What are the odds that, in this dark galaxy of high-security malware development, Flame is the most complex critter of all? That Flame has finally been detected, after running loose in the Middle East for two years, while all the other less-sophisticated worms - all of them - have remained undetected?

Almost nil, I'd say.

Flame just happened to get caught at a time when its discoverer, Kaspersky Lab, is in the middle of a global publicity campaign, and at a time when Iran finds it convenient politically to play the cyber victim card. Cue press releases.

So what's really going on out there?