Re: COBOL Compiler for Windows

Pete Dashwood wrote:
> OK. If I have this right, you are saying I was wrong to suggest they
will be 
> penetrated once they take over the role of "network server" (as some
people 
> suggested they will), because they've been connected to the network for 
> years already, and have already been penetrated?

More or less. In other words, I agree with you on what I think is the 
major thesis (mainframe systems will be attacked when they are 
accessible to attackers), though I disagree on the minor one (when and 
how rapidly this will happen).

I don't expect there will ever be an explosion of malware specifically 
for the mainframe OSes. They're a less-rewarding target (fewer of 
them, in particular) that requires greater effort. There's less 
interactive use, which means a smaller attack surface for trojans; and 
many of the worms these days are platform-independent, targeting 
multiplatform server languages (eg PHP) and subsystems (eg Oracle).

(Viruses are more difficult to write for any system with C2 or better 
security than they were for unsecured PC OSes, but they're also 
largely passé, and are much less useful to today's blackhats, most of 
whom are professional malware writers for organized crime and 
industrial espionage.)

The main security advantages of the mainframe over, say, Windows, are 
cultural. People don't tend to run unnecessary services on mainframes. 
People don't tend to use their mainframe servers for checking email 
and surfing the web. People who administer mainframes tend to be 
cranky about giving anyone permission to do anything.

Fact is, it's pretty rare these days that anything exciting happens in 
computer security. The arms race between attackers and defenders is 
fairly even, so attack and penetration rates stay fairly level over 
the long term. It's just noise to most people - until it's their 
system that's successfully penetrated.

-- 
Michael Wojcik