Good Worms Are a Bad Idea

Some bad ideas seem to live on forever. One of the big ones in computers is to use hacker tactics to perform white-hat operations on an Internet scale. The classic example of this is the "good worm" idea—a worm that spreads among computers to improve their security.

There have been attempts to do this in the past, most famously Welchia, a worm that exploited the infamous Windows RPC/DCOM network vulnerability in order to patch it. There were also reports years ago of Hewlett-Packard launching a good worm, but this turned out to be a more conventional scanning system that in some ways presaged NAC (Network Admission Control).

The lesson from Welchia is that these things are a bad idea. As Spencer Katt put it, good worms resemble their evil twins in many unpleasant ways. They invade privacy, hog resources and potentially create vulnerabilities in the systems they infect.

More to the point, releasing them is illegal and immoral. It's also pretentious: How dare you decide what code other people should be running on their systems? How dare you say that you know better how to manage my security than I do?

These issues are all in play in the recent seizure of a major botnet by researchers at TippingPoint Technologies' Digital Vaccine Laboratories. Even more, when an experienced group like TippingPoint is at the center, it's tempting to take some modest action—like killing known malicious processes. The researchers here, however, seem to know better.

David Endler of TippingPoint recognizes the dilemma and says that it is liability issues that made the decision clear. He's right in that liability issues are a great reason not to take action on botnets, but the moral ones are clear enough, too.

Efficacy is another potential reason. For many of the newly cleansed systems, good times will be short. The computers were vulnerable to begin with and probably many of the users would go right out and re-infect themselves. Unless the process could include applying patches, updating software and installing anti-malware, effectiveness would be less than we would hope for. For older systems—running Windows 98, for example—little can be done to make them secure enough.

Yet, it seems wrong to conclude from all my skepticism that we should just leave these machines alone, unfettered in their ability to attack innocent third parties. But if anyone should be taking bots offline without the explicit consent of the computer owner, it's a government, and bots are international. I certainly don't trust the United Nations to be mucking with people's computers, so I don't know where to go with this.

I do think someone will make a serious attempt to launch an "unbot" eventually. They'll have to do it surreptitiously, superhero-like, masking their secret identity. They could get away with it, too, but like other superheroes, they'll have to flout the law in order to fight crime, and innocent people will be damaged in the process. As in The Incredibles, expect lawsuits to follow. It's not a good situation.

Security Center Editor Larry Seltzer has worked in and written about the computer industry since 1983.

For insights on security coverage around the Web, take a look at eWEEK.com Security Center Editor Larry Seltzer's blog Cheap Hack