A warning to those who might try to stop the Storm worm: Be prepared for a counterattack.

Researchers say they have heard of several instances in which Storm -- the infamous botnet created by a widespread worm/Trojan distributed across the Internet -- has successfully launched reprisals against those who try to break it.

"This is the first time that I can remember ever seeing researchers who were actually afraid of an investigating an exploit," said Josh Corman, principal security strategist at IBM's Internet Security Systems unit, in an interview last week at Interop New York.

"The bad news about Storm is that it fights back," said Shane Coursen, senior technical consultant at Kaspersky Lab, during a session at the conference. "There have been cases in which a researcher was discovered, and within five seconds, he had a DDOS attack from 10,000 bots."

It's not clear whether Storm's reprisals are automated or manual, says Joe Stewart, a researcher at SecureWorks who has been studying Storm since it first emerged at the beginning of this year. "They could be triggered by behavior that's typical of a researcher, or they could be the result of someone studying logs and launching an attack in response," he says.

Storm can collect IP addresses of end users via HTTP and analyze them, Stewart notes. Researchers can use proxies, but if they are discovered, a DDOS counterattack might target the proxy, and innocent users might get hurt, he says.

Corman said that Storm's potential as a launching pad for DDOS attacks is enormous. "It hasn't been used that way yet, but the potential is there," he said. "If they wanted to take out a different financial institution every week, they could do it." Corman believes that estimates showing Storm at 50 million nodes are wildly inaccurate, but he believes that the botnet may be as big as 9 million nodes. (See Storm Worm Botnet Attacks Anti-Spam Firms.)

Stewart disagrees, stating that Storm "was never more than 1 million nodes," and likely is significantly smaller now that Microsoft and others have launched efforts to restrict its growth. "That's still more than enough to cause some serious problems," he notes.

Aside from potential reprisals, Storm worries researchers because it is a different sort of botnet that operates on a peer-to-peer basis, uses encryption, and is capable of synchronizing its attacks, Stewart says. "It works like a commercial product," he says. "If you were to study botnets and botnet creation in a textbook, this would be the one you would study."

Yet, despite its potential as a DDOS weapon, Stewart doesn't think Storm will frequently be used for such attacks. "I think it will continue to be used as it's being used now, for spam and pump-and-dump [stock] schemes," he says. "It wouldn't pay for the operators of Storm to do anything that would get them more attention, because then more resources will be turned to stop them. They want to keep operating under the radar."

Even if Storm's operators maintain a low profile, however, they will likely continue to see new researchers -- and take action against them. Storm has already been used on universities and anti-spam sites such as Spamhaus, according to reports. (See Storm Botnet Puts Up Defenses, Attacks Back.)

"If you're looking into it, be ready," said Corman.

Have a comment on this story? Please click "Discuss" below. If you'd like to contact Dark Reading's editors directly, send us a message.