Re: Hardware security metrics
From: Volker Birk (bumens_at_dingens.org)
Date: 09/15/05
- Previous message: mjr1007_at_yahoo.com: "Re: Hardware security metrics"
- In reply to: mjr1007_at_yahoo.com: "Re: Hardware security metrics"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Date: 15 Sep 2005 20:08:55 +0200
mjr1007@yahoo.com wrote:
> Just as there are performance metrics for hardware, (mips, megaflops,
> mips per watt...) I was hoping there might be similar metrics for
> security.
Usually, security has no metric. A metric is a relation, which fulfills
the triangle inequality. So it needs a measure, which can be quantified.
Security means, that you want to be safe from certain events.
To be safe from an event cannot be quantified, or single events can be
quantified with a probability only. So it is not trivial (if sensibly
possible at all) to define a security metric.
Usually, because of these facts people are classifiing bundles of
safety topics, so one can have something like a measure for security,
too: wether a system is of a predefined class or not, and classes
can be ordered.
I don't know such classifiing for hardware, though. I'd say, hardware
means that you have to trust the manufacturor. Of course, i.e. you can
proof VHDL code, too. But how do you want to guarantee, that what was
proofed, also is what was built?
Yours,
VB.
--
"Es kann nicht sein, dass die Frustrierten in Rom bestimmen, was in
deutschen Schlafzimmern passiert".
Harald Schmidt zum "Weltjugendtag"
- Previous message: mjr1007_at_yahoo.com: "Re: Hardware security metrics"
- In reply to: mjr1007_at_yahoo.com: "Re: Hardware security metrics"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Relevant Pages
|