From: "A. Padgett Peterson P.E. Information Security" <PADGETT@hobbes.orl.mmc.com>
Date: Thu, 30 Jan 1997 12:16:12 -0800 (PST)
To: jimbell@pacifier.com
Subject: Re: RC5-12/32/5 contest solved
Message-ID: <970130151428.20215ea4@hobbes.orl.mmc.com>
MIME-Version: 1.0
Content-Type: text/plain


Jim Bell rote:
>Assuming it's possible to build a chip which tests solutions in a 
>massively-pipelined mode, the 400,000 or so solutions per second tried (for 
>what is probably a $2000 machine) would probably increase to 100 million per 
>second per chip (at a cost of maybe $100 per chip, if implemented in 
>parallel).    That's 5000 times more economical,  which would translate to a 
>find in 2-3 days if the same dollars in hardware were invested.

Hi Jim, still on your list 8*)

Funny thing is that 3 1/2 hours for a 40 bit search is the "real world"
number I was using two years ago (can look it up in various archives)
so is interesting that the first real test came out exactly the same.

Is why I said 40 bits should not protect anything worth more than U$250.00

Have good reason to believe your estimate for a purpose built machine this
year (expect 600,000,000 to 1,000,000,000 kps per sieve - these will not 
be cheap chips but will be commecially available). Expect 400 arrays
would be required to do DES in a day (average) but is a lot more
achievable than the 65k postulated by the gang of nine.

Still would not be too concerned about using DES so long as every message 
encrypted (including orders for a tuna on rye) and each uses a different
key - is "security by obscurity" in a way but am comfortable with it.

Besides, if really concerned will just superencrypt.

					Warmly,
						Padgett

	"I love it when a plan comes together."