Viruses bring a sting in the tail
Published: 02 Jun 2005 14:00 BST
Fame — or infamy — just doesn't cut it any more for virus writers. If you get famous in this business then you're very likely to get arrested, and anyway there is now a much more attractive, green-backed, option out there. Computer Associates reported today that a compromised PC can be hired out for 3p a time to spammers and identity thieves, giving a real incentive to build botnets comprising thousands of compromised machines.
It's a change from the old days when the object of writing a virus was to have it spread as far as and fast as possible — and preferably further and faster than any that had gone before. Now the aim is to spread just enough but not too much.
Now imagine you are working for the other side, being paid to analyse the viruses and produce the patches. You have a limited amount of resources — people, time, and so on. So you turn to triage, figuring out which viruses pose the largest threat, and releasing patches for those first.
The result? A small number of viruses patched very quickly, but a very large number slipping through under the radar which don't get patched for days, weeks, or even longer. It is these that the industry should be worrying about. Each individual strain may only propagate so far and cause limited damage, but collectively they are likely to do a great deal of harm.
It's a new twist to the old long-tail principle that other industries are just beginning to wake up to, such as online book retailers who make half their total sales from books that individually sell only a couple of copies a month.
Email security outsourcing firms, who tend to use software from several security vendors, are well aware of the problem. Many have their own league tables showing just which vendors get the patches out first. And they can see that some viruses are being ignored. It is hard to see how antivirus vendors can effectively regain the lead here, if indeed they ever had it. But the growing botnet menace means we need a proactive approach from users and vendors alike, and that means building and selecting the most secure software possible.
In sales, the long tail can bring big rewards, but in the security industry it has nothing but a nasty sting.


