The multiple ways in which the Nimda worm is able to propagate makes it that much harder to defend against than other recent worms and viruses, security analysts said. But corporations that apply the latest Microsoft Corp. patches and use updated virus-protection software from antivirus vendors appear to be reasonably well protected against it.
Since the worm is capable of spreading via client PCs, corporations should also ensure that they disable the Java script functionality on end-user desktop machines and caution users about opening any unfamiliar e-mail, said Allen Householder, a member of Carnegie Mellon University's CERT Coordination Center.
And users will need to disconnect all infected servers from the network, reformat their hard drives and reinstall all the software from a secure source, Householder added.
"The newness of this is that it leverages a number of different vulnerabilities in order to propagate itself," Householder said.
The Nimda worm, reports of which first began flooding into mailing lists and security firms on Tuesday morning, is a mass-mailing piece of malicious code that infects systems running Microsoft Windows 95, 98, ME, NT and 2000.
The worm has been found to clog part of the Internet, slowing down or even stopping Web traffic for some users. Many sites are also experiencing high volumes of e-mail and network traffic as a result of this worm, according to a joint statement from CERT, the SANS Institute and the Information Technology Association of America. There appears to be no evidence the worm is linked to the recent terrorist attacks in New York and elsewhere, the statement said.
Unlike other worms and viruses, Nimda is capable of spreading via network-based e-mail as well as by Web browsers. It has also been tuned to look for and exploit backdoors left behind by previous viruses such as Code Red and Sadmind-II.
Based on initial analysis of the worm, Nimda appears to do little more than try and propagate itself via various means, including modifying Web content on infected Microsoft Web servers, Householder said.
What makes it dangerous is the various methods it uses to spread itself, making it necessary for users to protect themselves on multiple fronts, said Russ Cooper, editor of mailing list Windows NT BugTraq and an analyst at Reston, Virginia-based analyst firm TruSecure Corp.
"This is a very sophisticated worm with several new twists" Cooper said. "The combination of network-based and browser-based attacks, and the use of previously compromised hosts as delivery systems, shows a level of awareness" not previously seen, he said.
The Nimda worm is able to go from "client to client via e-mail, from client to client via open network shares, from Web server to client via browsing of compromised Web sites," according to a CERT advisory.
The worm also spreads from an infected client PC to Web servers running Microsoft's Internet Information Server (IIS). It does this by scanning the Internet for IIS servers that may have been compromised already by Code Red and other viruses and by exploiting a previously known vulnerability in IIS known as Unicode Directory Traversal, said Carey Nachenberg, chief researcher of Symantec Corp.'s security response team In fact it was a massive increase in such network scanning early Tuesday that first alerted network administrators of the worm.
In a four-hour period starting at approximately 9 a.m. EDT Tuesday, the University of Chicago's Web servers were scanned by nearly 7,000 unique IP addresses looking for vulnerabilities to exploit, said Larry Lidz, a senior network security officer at the university.
As a result of the attacks, about 20 university servers were infected with the Nimda worm and had to be disconnected from the network, Lidz said. Now, he is recommending that the infected systems be reformatted and all software reinstalled.
"If somebody has used a backdoor left by worms such as Code Red to infect your systems, you never really know what they have done to the system," Lidz said.
The worm also attacks via e-mail arriving as an apparently blank message with an attachment called "readme.exe." On vulnerable systems, the worm is triggered automatically by simply opening or previewing the mail message. Users can also trigger the worm by running the attachment, according to Cooper.
Once it infects a client system, the worm attempts to send itself to all the names in the address book. It also scans for and infects any vulnerable IIS server it can find on the network, Cooper added.
Once a server is infected, the worm searches for all directories on the server with any Web content, which it then proceeds to modify in such a manner as to facilitate further propagation. Vulnerable end users visiting such infected Web sites would infect their own systems without their knowledge, Householder said.
Another manner in which the worm spreads is via network shares. An infected client PC that is connected to a file server for instance, would infect files on that system as well. Under certain conditions, any users who attempt to then use those files would infect their PCs, CERT said.