, SecurityFocus 2005-10-12
Story continued from Page 1
Online criminals' use of bot nets to send spam and phishing e-mail messages convinced the U.S. Federal Trade Commission to kick off Operation Spam Zombies, a campaign focused on educating Internet service providers (ISPs) about what they could do to clean up the amount of compromised computers, also called zombies, on their networks.
"This particular project targets ISPs because we think they have a large ability to improve the situation where others could not," said Markus Heyder, a legal adviser in the FTC's Division of International Consumer Protection and the coordinator for the project."They are in the best position to understand the technology and secure their systems better than they are doing right now."
The FTC does not yet have data on whether its initiative has produced results, Heyder said.
While the recent successes by law enforcement have garnered some attention, the bot net problem will not likely be easily dismissed, security experts said in interviews this week.
For one, bot software that infects vulnerable computers has evolved and now typically consists of modular architectures into which new functionality can be plugged quickly and easily. For example, the latest exploits for Microsoft's operating system are incorporated into such bot software in weeks, if not days.
The three Dutch suspects are alleged to have used a customized version of the publicly available SDBot software to build their network of compromised computers. The software, dubbed W32.Toxbot by antivirus firms, records keystrokes and sends the information back to the attackers.
Such customization is standard, said Joe Stewart, a senior threat researcher for security firm LURHQ.
"Every group seems to take code and adapt it to their purposes," Stewart said.
Bot software is also taking its cues from the world of peer-to-peer networking. The latest software uses peer-to-peer technology to make it harder to shut down the network and to hide the location of the attacker, said Prolexic's Lyon.
Such innovations will likely continue because the crimes still reward the savvy bot programmer, Lyon said. In one case, for example, Prolexic was able to confirm that a group of bot herders had received more than $8 million from various extortion schemes. Generally, bot herders use their networks to threaten companies with a denial-of-service attack unless they pay. However, companies that do pay frequently find the attackers return to threaten them again and again.
"Eight million (dollars) is quite an incentive," Lyon said. "It also buys you a lot of new features for your bots from programmers."
Internet service providers have increasingly focused at heading off the problem by notifying customers on their networks that likely have computers compromised by bot software.
America Online, a division of Time Warner, has had to develop technology to battle bot nets, said Andrew Weinstein, a spokesman for Internet giant.
"We tend to notice DOS attacks and be able to respond to them faster than anyone," he said. "Technology has helped mitigate the most severe affects of DDOS attacks."
Not all the efforts of ISPs work well, however. Some providers block Internet Relay Chat (IRC), a common protocol for controlling bot nets, but many of today's bots use nonstandard channels, or ports, to communicate, said LURHQ's Stewart. I
It's an arms race in which both the Internet service providers and law enforcement have to work harder to get ahead, he said.
"People making money off of it are not going to stop because someone else in a different country got arrested or because a large bot net got taken down," Stewart said. "Hopefully, as law enforcement gets more clued in to how bot nets operate, we will get a critical mass where it acts as an actual deterrent to these people."