At the same time, there is an exploit out in the wild that performs a distributed denial-of-service by crashing the attacked system. DDoS attacks are a bad thing, of course, but they arent as much of a worry from a mass-attack standpoint. Authors cant make a worm out of a DDoS attack because if the system crashes, theres scant opportunity to trick the owner into spreading the worm.
A real worm requires a means of infection and the ability to execute arbitrary code on the infected system. The Microsoft advisory indicates that this is possible with the ASN.1 issue.
There have been allegations that the claim of arbitrary code execution is an exaggeration, however, experts advised me that a code execution worm is merely difficult, but not impossible. Given a large number of vulnerable systems in the world, such a worm could still spread.
Heres what its all about: Without getting too specific, its not possible to write a reliable exploit that gains control of the target system.
According to Ken Dunham, director of malicious code at iDEFENSE, "The heap overflow exploit is proving to be more difficult for some attackers than what they had originally thought. A change of state makes it more difficult to successfully exploit computers."
In other words, the attack requires conditions on the PC which cannot be predicted or controlled by the attacker.
But even if, for the sake of argument, only 1 percent of attacks succeed, that would encompass enough computers for the worm to spread far, assuming it wont do serious damage to systems that it cant successfully infect. The worm will keep trying and trying and eventually, it will get through— if there are enough unpatched systems for it to find.
Meanwhile, if unsuccessful attacks crash the system or do something else to tip the owner off of the worms presence, the jig will soon be up.