Re: [secdir] [Cfrg] ISE seeks help with some crypto drafts

[Aaron Zauner <[e-mail]>]

Hi David,

> On 09.03.2019, at 14:25, mcgrew <[e-mail]> wrote:
> 
> Hi Aaron,
> 
> The IPR statements for your draft seem to only cover the use of OCB in TLS, and not its use outside that context.   I am not a lawyer, but that is how I would understand the phrase “a royalty-free non-exclusive license to all claims of the referenced patents needed to realize a fully compliant implementation of TLS (Transport Layer Security) supporting AES-OCB (RFC 7253)”.   This point might not matter to the implementers of your draft, but it might matter a lot to other people.  

That’s correct. My bad, the initial text was supposed to say "general IPR exemption for OCB(3) in TLS for IETF”. I wrote & edited that mail right after getting up and must have removed that editing this message and moving sentences around in an attempt to make it more coherent :)

> 
> For anyone researching the IPR status of RFC 7253, please note that Phil submitted several IPR statements IPR statements related to (the draft version of) that specification, which can be found with this search: https://datatracker.ietf.org/ipr/search/?draft=&rfc=7253&submit=rfc&doctitle=&group=&holder=&iprtitle=&patent=    

Yes, you can find all of Rogaway’s IPR claims pertaining IETF documents via: https://datatracker.ietf.org/ipr/search/?draft=&rfc=&doctitle=&group=&holder=Phillip+Rogaway&submit=holder&iprtitle=&patent=

> Aaron, let me also take the time to thank you for working with the patent holders to get the IPR statements issued for your draft (and thanks are due to the patent holders as well).   

It was quite an interesting experience and I hope that writing it up can serve others in the future. My impression is that it’s not difficult to get an IPR exemption if the people behind the original work are intrested in it being used. I think Rogaway had just the best intentions at heart with the original licensing, but may have missed that it’ll cause issues with standardization processes in the future/real world that may have an adverse impact on the use of OCB3. He’s granted free licenses to Open-source implementations and other projects and organizations since. OCB3 (and my draft) were also implemented in various libraries, among them OpenSSL (C), Botan (C++) and BouncyCastle (Java). Back when they optimized it OCB even beat GCM w.r.t. performance in OpenSSL on Haswell Intel x86_64 processors and has since kept up pretty well consindering AESNI is aimed at GCM/GHASH (see head comment on performance in https://github.com/openssl/openssl/blob/master/crypto/aes/asm/aesni-x86_64.pl).

```
  *) Added support for OCB mode. OpenSSL has been granted a patent license
     compatible with the OpenSSL license for use of OCB. Details are available
     at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
     for OCB can be removed by calling config with no-ocb.
     [Matt Caswell]
``` — https://www.openssl.org/news/cl111.txt (in "Changes between 1.0.2h and 1.1.0  [25 Aug 2016]”)

Aaron