Conficker's 6m strong botnet confounds security probes
Cross-industry group baffled by super crypto shield of infamous botnet
The Conficker Working Group constantly monitors the IP addresses of infected machines as they check into sink holes. Many enterprises associated with infections drop off the radar only to return days or weeks later, probably as the result of the application of infected backups that have not been purged of malware.
Utilities such as Microsoft's Malicious Software Removal Tool, effective in cleaning up other infections, have proved ineffective against Conficker because software security updates get disabled on compromised machines.
"Conficker has created a stable pool of around six million infected machines," Joffe warned. "These machines are running without updates or anti-virus defences which creates a wide open backdoor."
Joffe warns that criminal hackers could register "phone home" domains associated with Conficker to harvest a list of infected IP addresses. This list could then be used as a targeting mechanism for attacks, safe in the knowledge that machines as the receiving end of exploits would be defenceless.
Attacks could be narrowed down to target systems known to be insecure rather than scanning for vulnerabilities, a noisy activity likely to set off warnings.
Blocking the registration of domains would thwart such an approach, but a "number of registries" are no longer collaborating with the Conficker Working Group, Joffe warned.
The miscreants behind the worm remain unknown and the purpose of the malware unclear. Joffe would only say that there was an active law enforcement investigation without giving any clues as to how it was proceeding.
Joffe said the cybercrooks behind Conficker were highly sophisticated and competent coders who used "bleeding edge" cryptographic techniques and responded to the efforts of the anti-virus community with progressively more sophisticated variants of their malware.
The latest variants of Conficker use MD6, a candidate for a federal encryption standard due to be formalised only in 2012, and the crooks were smart enough to apply patches days after earlier shortcomings of the scheme were discovered by cryptoanalysts.
"We can't take over the botnet and we don't know what it's doing - it's uncrackable," Joffe explained. "The cybercrooks behind the Conficker might hive off components of the botnet to other bad guys and we wouldn't have a clue. The loss of hundreds of machines would be indistinguishable from background noise."
Although perhaps not as active as during its peak last year, the core part of the committee remains active and meets every fortnight. Activities including outreach to hospitals (many of which were hit by serious Conficker infections only this year) remain an important part of its work. "We'll continue as long as the Conficker threat remains," Joffe concluded, adding that 18 months after the first outbreak new infections are still appearing.
The Conficker Working Group, the first cross-industry organisation of its kind to fight a malware threat, formed the blueprint for the later Mariposa Working Group. The takedown of the main 12 million strong information-stealing Mariposa botnet and the arrests of three suspects in Spain prompted a decision to disband this group, whose work had reached completion.
An active law investigation into the trio in Spain, alleged clients of the Butterfly botnet client allegedly developed by Slovenian coders, remains ongoing. A suspect, later identified as Dejan Janžekovic by ex-Washington Post staffer Brian Krebs, has become the focus of the police investigation in Slovenia.
Joffe declined to answer questions about the Mariposa investigation beyond saying that two or three other groups who used the Butterfly botnet kit are of potential interest to international police investigating the case. "The Spanish group were very successful in getting the Butterfly botnet kit, but its use was fairly widespread." ®
Sponsored: Network DDoS protection