Re: [oc] crypto-accelerator

[tom st denis <tomstdenis@yahoo.com>]

--- Wamnet@aol.com wrote:
> In a message dated 9/11/2000 7:38:55 AM Eastern Daylight Time, 
> tomstdenis@yahoo.com writes:
> 
> > --- Paulus M Tamba <ulus132@cyberlib.itb.ac.id> wrote:
> >  > hi all
> >  > 
> >  > i would like to work on crpyto accelerator, but i couldn't find
> any
> >  > website which gives brief explanation about crypto. i would like
> to
> >  > ask
> >  > you all where to find about crypto.
> >  
> >  Well what do you want to implement?  A general purpose Crypto
> >  Accelerator should have
> >  
> >  a) A Block cipher, preferably AES (not yet decided)
> >  b) A Hash Function, preferably SHA-1
> >  c) A Large Integer ALU, such as a 1024-bit ALU
> >  d) A physical RNG device (thermal noise, etc...)
> >  
> >  If the device has all four a-d components it would be very popular
> and
> >  usefull.  It should also have some memory to hold keys on the
> device,
> >  so possibly it would need a few kb of ram...
> >  
> >  Then it should have a nice PCI or ISA interface say using a few
> memory
> >  locations (memory mapped io) or actual I/O ports.  I can provide
> >  analysis and support, but I haven't much clues about designing
> >  hardware.
> >  
> >  Tom
> >  
> 
> 
> A German site published an entire open source VHDL DES design. I have
> the 
> source if anyone wants it. We've found that Broadcom and Hi/Fn
> already have 
> some pretty good stuff working 3DES/SHA-1 into the Gigabit speeds so
> we kinda 
> shelved the open source stuff. 

If you can do it an ounce faster then a x86, and it's cheap who cares
about gigabit?  I would use it just to accelerate day-to-day usage of
say PGP.

Tom

__________________________________________________
Do You Yahoo!?
Yahoo! Mail - Free email you can access from anywhere!
http://mail.yahoo.com/