Just a few years ago Windows users, even responsible Windows users, had good reason to be fearful of the attack that would slip past their defenses or their notice.
Things have changed. Nobody should ever be complacent, but a responsible user can be confident that defensive software and good habits will protect them. More interestingly, attacks just arent what they used to be.
A report by Alexander Gostev, senior virus analyst at Kaspersky Lab, indicates that innovation in malware development is stagnant. There have been no major developments in some time. In fact, there have been no major attacks since the release of Zotob in August 2005.
Zotob, incidentally, targeted mainly Windows 2000 systems and XP SP1 to a lesser degree. What Microsoft has been saying about XP SP2 is true: Users are much safer running XP SP2 than earlier versions of Windows. Their own data from their Malicious Software Removal Tool (Word .doc file) shows as much, and in fact probably understates the matter.
There have been a number of small attacks. Some of them, like the WMF vulnerability, enter the background of the malware scene and will be with us for a long time. Perhaps the most prominent security term of 2006 was "targeted attack." We had quite a few of them, mostly centered around zero-day vulnerabilities in Microsoft Office. See the Kaspersky report for more interesting details on these vulnerabilities.
The focus on vulnerabilities generally is another point in the report. As I said, there is no innovation anymore in malware—except where it involves the exploit of a vulnerability, especially a zero-day exploit. But even these are often less of a threat than they used to be. A few years ago vulnerabilities brought us attacks like Blaster and Sasser, where users could be infected over the Internet while they were asleep. Now the exploit usually involves substantial user action and can often be blocked by anti-virus software.
Now Im going to re-ask a question Ive been asking vendors for a while now without what I consider a full answer: How many truly new infections do we have these days as opposed to re-compromises of systems already infected with other malware? I think the latter is where the action has been for a long time now. My theory that a very large percentage of new infections are on already-infected systems has two main arguments in its favor:
- The systems are compromised, providing a hole through which new malware can attack
- The users are proven to be willing to click on things they shouldnt
When you look at the data from AV companies on what the most prevalent attacks are its like a trip through the way-back machine. Look at Sophos report on the Top 10 viruses reported to Sophos in October 2006, and bear in mind that Sophos doesnt have much of a consumer presence, just business: