Re: SRP in OpenSSH draft protocol spec

[nisse%lysator.liu.se@localhost (Niels Möller)]

RJ Atkinson <rja%inet.org@localhost> writes:

>         Does anyone know of any patents pending or patents
> granted that cover SRP ?  My possibly incorrect understanding
> is that SRP is patented or patent-pending.

It's patented by Stanford, and the basic variant (that described in
the SRP rfc) is licensed freely. See http://www.ietf.org/ietf/IPR/WU-SRP:

: Received December 22, 2000
: From: Thomas Wu <tjw%CS.Stanford.EDU@localhost>
: 
: The SRP Authentication and Key Exchange System, as specified in
: RFC 2945, is available royalty-free worldwide for commercial and
: non-commercial use.
: 
: Extended variants of SRP, such as those based on SRP-Z, may require
: a license, which Stanford will grant on a non-exclusive basis, under
: reasonable and non-discriminatory terms.
: 
: For questions about SRP, please contact me or visit
: http://otl.stanford.edu/
: 
: Tom Wu
: tjw%CS.Stanford.EDU@localhost

I've discussed this some more with Tom Wu. I'm not sure it would be
appropriate to quote it all here, but the main point seem to be that
SRP is patented, but licensed freely. Thus, the SRP case is not much
different from CAST-128 and CAST-256, patented by Entrust, or for that
matter DES, which I believe was patented (by IBM?) and then freely
licensed after it became a US government standard.

Tom Wu wrote:

"Stanford's primary intention is to keep the strong password field
 open, by making a well-tested, well-standardized method available for
 free. It simply wouldn't make sense for us to change our minds about
 licensing, because it would undermine a great deal of work, time, and
 effort that we've already put in; that's why it won't happen."

I'm by no means a patent expert, but I feel that SRP is free enough
for use in free software, and that ought to be free enough for
anybody.

It would be good if someone more official at stanford could back up
Tom's statement, though.

/Niels