[cap-talk] High level dissonance

[Ivan Krstić]

On Feb 23, 2008, at 7:10 PM, Karp, Alan H wrote:
> It's a bit general because there are some attacks that are not on  
> users, e.g., some viruses.

I submit anyone making this statement is thinking about the problem  
incorrectly, or at the very least, not sufficiently thoroughly.

Every computer attack is an attack on the user. If it weren't so, we  
wouldn't much care about the attack. Viruses that destroy data are an  
attack on the user's data; viruses that render the machine inoperable  
are an attack on the user's ability to get her work done. Viruses and  
trojans that zombify a machine and join a botnet are an attack on  
_some other_ user's ability to get her work done.

We can differentiate between malware that harms the user of the  
machine it infects and that which harms the user of some other  
machine, but ultimately it's _all_ about the user. This goes for  
understanding defense as well as offense. The reasons viruses never  
took off widely on UNIX is because the software installation model is  
different: most software is installed (and owned, in terms of file  
permissions) by the administrator, not the individual user. Viruses  
executing with the user's permissions thus rarely have binaries to  
infect. On Windows, this wasn't the case; the model had users in  
charge of installing software because it provided _the user_ the  
convenience of not having to switch to a different account to perform  
those actions. If this wasn't the case, Windows could simply prohibit  
any non-system binaries from modifying any other binaries or DLLs and  
the concept of viruses would, overnight, cease to work.

--
Ivan Krstić <krstic at solarsail.hcs.harvard.edu> | http://radian.org