A hacker has posted code that could be used to take over a system running Microsoft IIS (Internet Information Services) server.
The software, which was posted to the Milw0rm Web site on Monday, could be a big problem for some webmasters, however the attack appears to work only on older versions of Microsoft's products.
It was not immediately clear how many versions of Microsoft's products are vulnerable to the attack, and Microsoft has not yet posted a security advisory on the issue.
The flaw lies in the File Transfer Protocol (FTP) software used by IIS to move large files around the Internet, so the victim would have to have FTP enabled in order to be vulnerable to the attack.
According to the Milw0rm post, an attacker could use this code to install unauthorized software on the server.
According to Nikolaos Rangos, the Milw0rm poster, the code works on Microsoft's decade-old Windows 2000 operating system, while running the older IIS 5.0 server. For the attack to work, the hacker would also need to be able to create a directory on the server, security experts say.
Other versions of IIS are also at risk, according to Thierry Zoller, chief security officer with security consultancy G-SEC, who has studied the issue. However, newer versions of Microsoft's operating systems have features that make it less serious, he added via instant message.
Rangos, who discovered the flaw, said he scanned several ranges of IP addresses to see how many IIS servers running FTP he could find. He said he found a lot of Windows Server 2003 systems running the FTP software with IIS 6. However, IIS 5 on Windows 2000 is "vanishing slowly," he said.
The attack also works on IIS 6, he said, but because IIS 6 has a defensive measure called "stack cookie protection," the code can only be used to crash IIS 6 running on Windows Server 2003, Rangos said.
Rangos did not notify Microsoft before posting his attack code, so the company has not had much time to work out a fix.
In an e-mail message, Microsoft said it was looking into the matter and was "unaware of any attacks trying to use the claimed vulnerability or of customer impact."