, SecurityFocus 2000-10-30
When Microsoft kicked a hacker off its network, the FBI may have lost its chance for a bust.
“
If you want to track somebody... you watch them very carefully without giving them any kind of tip-off. That's how I was caught.
”
Microsoft said over the weekend that the hacker only had access for twelve days, and that corporate security agents monitored the outlaw for most of that time. The intruder found some of Microsoft's proprietary source code, but did not download any of it, the company says. According to a New York Times report, Microsoft finally called in the FBI on Wednesday, October 25th, but not before closing down all of the intruder's purloined accounts, effectively barring him or her from the network.
That move may have staunched the hemorrhaging of Microsoft's corporate secrets, but it also eliminated the FBI's opportunity to catch the intruder with his or her hand in the cookie jar. "Unless Microsoft was able to give them a lot of information, they're going to have to start from scratch," says Eric Smith, a former Air Force computer crime investigator who helped track the teenaged culprits in the 1998 "Solar Sunrise" Pentagon intrusions, and later traced Max Butler, a California intrusion detection expert who plead guilty last month to cracking Defense Department systems in his spare time. "They aren't going to be able to monitor things, because Microsoft shut him off," says Smith.
The choice between trapping and evicting an intruder is a recurring dilemma in cyber security. Internet outlaws typically hide behind a daisy chain of hijacked systems. Using subpoenas and trap-and-trace warrants, FBI investigators can track an attacker through those systems while the intrusion is underway. But once an intruder has been kicked off, or spooked into abandoning an attack, agents are left chasing a fading contrail of connection logs, which a canny hacker might very well modify or erase. "It's a lot harder," says Smith, "because you never know if the logs are, in fact, what they actually appear to be."
Consequently, some of the biggest hacker cases have been solved when the victims allowed an intruder to continue working unmolested, right up until the moment law enforcement agents knocked on the suspect's door.
Perhaps most famously, astronomer Cliff Stoll allowed a hacker to use a Lawrence Berkeley Lab machine as a platform to launch attacks on dozens of military sites and defense contractors throughout the summer of 1986, while privately warning the other victims and driving an investigation that ultimately resulted in the arrest of "Hanover Hacker" Markus Hess in Germany. In 1995, similar tactics snared hacker Kevin Mitnick in his Raleigh, North Carolina hideout.
"If you want to track somebody and they have access to a box where they can't do any damage, than you watch them very carefully without giving them any kind of tip-off," says Mitnick. "That's how I was caught."
But few organizations are willing to accept the risks of allowing an attack to proceed in exchange for the nebulous benefits of bringing about an intruder's arrest. With the FBI struggling to persuade victim companies to report incidents in the first place, Microsoft's decision to lock its virtual doors before phoning the bureau is overshadowed by the fact that it made the call at all.
"If he actually had access to the source code, I'd probably have cut my losses and shut him out," says Smith, now a security consultant with Colorado-based
And the odds of convicting the mysterious Microsoft hacker, or hackers, now that the network is closed and the entire matter is an international news story? Not good, says Smith. "Once it hits the press, the first thing those guys are going to do is go to their computer and wipe out everything."
