The CERT Coordination Center has received reports of widespread attacks using a recently disclosed security vulnerability and a previously unknown security hole in Microsoft's Windows operating system.

The known vulnerability is in a Windows interface that handles communications using the RPC (Remote Procedure Call) protocol. That problem affects most supported versions of the Windows operating system and could enable remote attackers to gain total control over vulnerable systems.

Microsoft disclosed the problem and provided a patch for vulnerable systems in a security bulletin July 16 after it was discovered and reported to the company by a Polish security group, Last Stage of Delirium.

Scan move Reports of Internet scans for vulnerable systems and isolated attacks using the Windows flaw surfaced after the publication on the Internet last week of code that could be used to exploit vulnerable systems, prompting warnings from security experts and the US Department of Homeland Security about pending attacks.

CERT's advisory cites the existence of multiple exploits on the Internet and reports of "intruders actively scanning for and exploiting" the Microsoft vulnerability.

In the past two days, the computer security research and development organization has received reports of "thousands" of systems compromised using variations of the malicious code, known as DCOM RPC, after the flawed Windows DCOM (distributed component object model) interface, according to Jeff Havrilla, Internet security analyst at CERT in Pittsburgh.

Hack estimate CERT does not know how many machines have been hacked using the DCOM RPC exploit, but the ratio between the number of reports CERT typically receives to the number of machines compromised suggests that the number is large, Havrilla said.

In many cases, administrators may not even know their computers have been compromised, he said. That may be by design.

In contrast to previous exploits such as the Code Red and Slammer worms, both of which took advantage of widespread Microsoft security holes, malicious hackers are keeping a low profile with the DCOM RPC vulnerability, Havrilla said.

"It's not like Code Red where the machine that was taken over started to spawn hundreds of processes. These machines are much more stealthy and much more controlled," he said.

Malicious Malicious hackers are using DCOM RPC to place Trojan programs on compromised systems, giving them control of those systems for use in future attacks, he said.

"It's not a worm. It doesn't rise to that level. But it's a controlled, systematic exploitation of vulnerable systems," Havrilla said.

Also adding to volume of attacks is a previously unknown flaw, also in the Windows RPC component but not covered by the recent Microsoft patch, that makes Windows 2000 machines vulnerable to denial of service attacks, he said.

That flaw surfaced in recent days, along with code to exploit the vulnerability, and has also been used in attacks on vulnerable systems. Those attacks have caused confusion among Microsoft customers who applied the RPC patch to Windows 2000 systems, fueling speculation that the Redmond, Washington, company's patch did not work.

In its advisory, CERT recommends that customers apply both the Microsoft patch and block network traffic on Transmission Control Protocol (TCP) ports 135, 139 and 445, which are used by the RPC service.