In a coderpunks thread entitled, "Debunking rumours about Rijndael," Marcus
Watts <mdwumich.edu> wrote:

<snip>
> It's mildly disappointing the versions mentioned here
> http://home.cyber.ee/helger/implementations/
> are only commercially available; I'm seriously thinking of teaching
> kerberos 5 how to talk rijndael (I already have it talking cast5 & rc6)
> but this is obviously only useful with code that I can freely
> redistribute. <snip>

Hi Marcus,

Are you aware that Rivest's RC6 is patented? That RSA is actively licensing
it for OEM and enterprise apps? FYI, I'll attach below what used to be
RSA's boilerplate reply to questions about its RC6 IP.

(If you want to explore this further, you should get T&Cs from the horse's
mouth. I don't work for RSA, and I certainly don't speak for RSA; although
I was a consultant to the company for many years.)

RSA did agree to make RC6 available, royalty-free, if the cipher was chosen
as the AES. Since Rijndael won the brass ring, however, RSA chose to
maintain its exclusive IP rights on Ron's RC6.

(One of the RC5 patents gives them an IP beachhead already, but there are
additional RC6 patent applications pending, in both the US and abroad.)

Needless to say, however, there are _lots_ of RC6 implementations floating
around in the aftermath of the AES evaluation process. (I recall that you
implemented both RC6 and CAST5 for K1.0 awhile back. Did anyone, btw, ever
point out Jeff Altman's RFC2949 as a solution to your Telnet/CAST problem?)

Your comments to the List about Kerboros' crypto seemed to imply that your
RC6 code is unencumbered and that you can "freely redistribute" it.
(Although I worry, silly me, that I'm tossing raw meat to a pride of
lions), I thought a pointer to RSA's IP claims might be useful.

Carlisle Adams' CAST5-128 was also patented, of course, but in 1997 Nortel
and Entrust made it available for all on a royalty-free
basis. Unencumbered, CAST128 was subsequently embraced by the S/MIME,
OpenPGP, and IPsec standards groups, and many developers, both commercial
and non-commercial.

Ron Rivest and RSA took a different path, as is their right. They never
released Rivest's RC6 into the wild. (Patents, as I am sure you know,
allow for publication without any sacrifice of IP rights.)

Of course, for all I know, RSA might be willing to cut some special deal
for you and the Kerberos Klan if RC6's simplicity and strength makes it, in
some way I don't understand, uniquely appropriate for use in Kerberos --
but I wouldn't push too far into any major distribution plan without coming
to an agreement with the boys and girls in Bedford.

Suerte,

         _Vin

------------------------------------------
[RC6 Boilerplate, from RSA's OEM and Developer Solutions Group.]

> RSA Security holds the following US Patent on the RC6(R) algorithm:
>
>"Enhanced Block Ciphers with Data-Dependent Rotations," U.S. Patent No.
>6,269,163.
>
>RSA Security also has a pending US patent application and several pending
>foreign applications on the RC6 algorithm. Thus, third parties may not sell
>or offer to sell in the U.S. implementations of the RC6 algorithm without a
>license from RSA Security. In addition, the mark "RC6" is a registered
>trademark of RSA Security and may not be used by third parties creating
>implementations of the algorithm without a license from RSA Security.
>
>RSA Security also maintains copyrights and other intellectual property
>rights in RSA Security's own implementations of the RC6 algorithm, which are
>incorporated into RSA Security products and made available to other
>organizations under license.
>
> RSA Security is willing, upon request, to grant non-exclusive
>licenses to its implementations of the RC6 algorithm on a nondiscriminatory
>basis and on reasonable terms and conditions.

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]