From the NewScientist story:
Because no central server needs to provide and coordinate all the downloads, software patches that spread like worms could be faster and easier to distribute because no central server must bear all the load. "These strategies can minimise the amount of global traffic across the network,"
I'm sure it would work. Except for those times that it doesn't.
Maybe on subnets of corporate networks where each system image is the same, this might be a safe way to deploy patches after they've been tested.
But what happens if a visitor's notebook is patched by this "friendly" worm -- and the patch crashes this visitor's system and results in significant data loss? Who is responsible for that mess? (Hey, I just wanted to check my e-mail!)
And this would be a really bad idea if ISPs decided to start dropping worms that would infect and then patch their subscribers. I want the ability to establish a restore point on my Windows images before I patch. And I like running a SuperDuper! backup on both my MacBook Pro and Mac Pro before I patch Leopard. That way, if something bad happens during -- or because of -- the update I can always make my world good again.
For those security pros that have been around awhile, they'll recognize that this "friendly" worm idea isn't new. In fact, there was a lot of talk about "Code Green" type worms to patch systems vulnerable to Code Red infections back in the very "wormy" summer and fall of 2001.
And in the 1970s, a virus known as Creeper struck the Advanced Research Projects Agency Network (the precursor to what is now the Internet). It displayed something to the effect of "Catch me if you can" as it spread.
Someone tried exactly that, with a program dubbed Reaper. To the best of my understanding, the Reaper did hunt down and kill Creeper without any negative impact. But it was a simpler tech world back then. Today, no CISO would ever want an application they couldn't control hopping around their subnets sniffing for things to patch.
I think the risks aren't worth the reward. And that these friendly worms could just as easily be rigged to inflict harm. And then there's the Law of Unintended Consequences that one of these worms won't quite work as planned. So what do you think? Would it be a good idea if someone wrote a worm named "Clear Skies" to go out and hunt down Storm botnets?