home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky sci.crypt:6385 alt.society.civil-liberty:7248
- Newsgroups: sci.crypt,alt.society.civil-liberty
- Path: sparky!uunet!wupost!uwm.edu!linac!att!att!dptg!ulysses!ulysses!smb
- From: smb@research.att.com (Steven Bellovin)
- Subject: Re: Question from someone who's new to all this
- Message-ID: <1993Jan5.005617.7945@ulysses.att.com>
- Date: Tue, 5 Jan 1993 00:56:17 GMT
- References: <1993Jan3.165050.27910@zip.eecs.umich.edu> <1993Jan3.170453.28445@zip.eecs.umich.edu> <bontchev.726158979@fbihh> <1993Jan4.145637.5551@newstand.syr.edu>
- Organization: AT&T Bell Laboratories
- Lines: 53
-
- In article <1993Jan4.145637.5551@newstand.syr.edu>, greeny@top.cis.syr.edu (J. S. Greenfield) writes:
- > This sounds far too broad. I would guess it's far more specific--somewhere
- > along the lines of use of modular exponentiation with a modulus
- > consisting of the product of two primes for cryptographic purposes.
- >
- > Modular arithmetic was not at all new to cryptography when RSA was
- > developed, so the patent would *have* to be much narrower than just that.
-
- You're right. Here's Claim 1 of the RSA patent:
-
- 1. A cryptographic communications system comprising:
- A. a communications channel,
- B. an encoding means coupled to said channel and adapted for
- transforming a transmit message word signal M to a ciphertext word
- signal C and for transmitting C on said channel,
- where M corresponds to a number representative of a message and
-
- 0 <= M <= n-1
-
- where n is a composite number of the form
-
- n = p.q
-
- where p and q are prime numbers, and
- where C corresponds to a number representative of an enciphered
- form of said message and corresponds to
-
- C == M^e (mod n)
-
- where e is a number relatively prime to lcm(p-1, q-1), and
- C. a decoding means coupled to said channel and adapted for receiving
- C from said channel and for transforming C to a receive message
- word signal M'
- where M' corresponds to a number representative of a deciphered
- form of C and corresponds to
-
- M' == C^d (mod n)
-
- where d is a multiplicative inverse of e (mod lcm(p-1, q-1)).
-
- It's very specific -- to infringe this claim, you'd have to use a device
- with all three parts. (N.B. -- before you start picking this apart,
- and looking for the obvious ways around it, note two things. First,
- the RSA patent has about 40 claims, each covering a different variation.
- Second, variations that are ``obvious to one skilled in the art'' also
- infringe.)
-
- Note, too, that RSA did *not* patent the equations themselves, nor did
- they patent an algorithm. They patented a ``cryptographic communications
- system'' whose behavior is described by certain equations. One way to
- look at it -- in fact, the way I look at it -- is to regard this as
- equivalent to patenting an electronic circuit whose behavior is described
- by certain other equations -- as, of course, is generally the case.
-