Microsoft’s HoneyMonkeys

Microsoft has announced a new security programme that uses automated ‘HoneyMonkeys’ to patrol the Web and seek sites that automatically install malicious code on Windows XP systems.

In its first month, the company’s project, named ‘Strider HoneyMonkey research project,’ located 752 Web addresses linking to 287 sites that could automatically infect unpatched machines. The project also discovered an attack that could penetrate a fully updated Windows XP Service Pack 2 system using a previously unknown vulnerability. Microsoft first discussed the HoneyMonkey programme in May and published a research paper discussing the details.

The system uses a chain of HoneyMonkeys, a name derived from ‘honeypots,’ which refers to passive security research server systems set up to wait for attacks. Each HoneyMonkey is a Windows XP system with a different level of patching running on a virtual machine. An initial wave of unpatched HoneyMonkeys scours the Web seeking potentially malicious sites. When a site is found that installs potentially malicious code, the virtual machine is scrapped and another takes its place.

The target URL is then passed to a virtual machine with a greater level of patching to see which systems are vulnerable to the site’s exploit. At the end of the chain is a fully patched Windows XP system, Microsoft said. The system builds a topology graph based on traffic redirection, which has led to the identification of a few major players who are responsible for a large number of exploit pages.

The project is relatively limited in scope. It only looks for code that can be installed with no user interaction, leaving out the more sophisticated and increasingly successful attacks relying on social engineering—attacks such as phishing. However, Microsoft believes the automated approach could become a valuable tool for detecting new types of attacks before they become widespread.