In the heat of a malware outbreak there is usually a lot of confusion about what variant of what worm is involved? Is it just a new variant or a completely new worm?
Inconsistencies between vendors about variant indices and virus names add to the confusion.
The latest effort to address this problem is the CME (Common Malware Enumeration) Initiative from a company called MITRE but sponsored by US-Cert. These are the same people who brought us CVE (Common Vulnerabilities and Exposures), the project on which CME is based.
The idea is to assign a specific identifier to each malware implementation.
To illustrate the problem, consider the table below containing data from AV-Test, a research project at the Otto-von-Guericke University Magdeburg (Germany).
What CME Is Up
This data is from a recent Bagle outbreak, and most if the identifications are of Bagle, but theyre almost all different specific variants.
Many (such as Ikarus) use a generic identifier.
Think of the confusion this might cause, and specifically look at the three big anti-virus companies: McAfee, Symantec and Trend Micro use names bearing no resemblance to each other.
Theres definitely a problem here, and its not an inconsequential one. But Cert is being restrained in their claims about CME and they should be.
Its not going to solve the problem.
CVE is much more successful, although its basically a behind-the-scenes system.
It is used to identify vulnerabilities. Frequently, if not most of the time, it is the vendor of a product who works with MITRE to assign a CVE candidate number to a vulnerability.
Vulnerabilities are usually disclosed to the world by one researcher or firm and accompanied by a definition of their nature. Malware is different.
Malware isnt disclosed, it is released to the world without announcement.
In the first hours, different anti-virus companies will be getting copies, in some cases arguing (as the table above illustrates) what exactly the threat is.
A procedure has been defined by CME wherein submissions will be sent and a candidate number assigned.
There are checks in place to try to avoid duplicate submissions and subscribing anti-virus companies are expected to use the ID number, perhaps in addition to their own name, in all correspondence.
If this works it will be great, but theres reason to believe that it will end up being just another name to look at.
Anti-virus companies already share code samples extensively through back channels, and still there are often disagreements during outbreaks about whether new threats are really new.
Id be shocked if the CME system supplants the more fun naming systems in place.
Part of the glory a researcher gets for finding and identifying a new malware is to name it, and they wont give that up easily.
Plus, the numbering system doesnt convey the existence of families of malware, such as Bagle, and the relationships between them.
You might assume that Sober.A and Sober.B are closely related, but are CME-123 and CME-124? Not necessarily.
Finally, having a single identifier in place, if the system would be really successful, would help make it clearer which anti-virus companies react faster and better than others, although AV-Test already has good testing to make that clear.
Such clarity isnt necessarily in the interest of the largest companies in the business.
I seem to be saying this a lot lately: I dont have a better idea; I think the problem is basically insoluble.
If were lucky, CME can gain general acceptance, and that will make things somewhat clearer, but I worry that too many factors are working against it.
Security Center Editor Larry Seltzer has worked in and written about the computer industry since 1983. He can be reached at [email protected]