[rabbitmq-discuss] RedHat's Pending AMQP w/ XML Patent

[Alexis Richardson]

Matt,

On Mon, Mar 16, 2009 at 3:23 PM, Matt Heitzenroder
<mheitzenroder at gmail.com> wrote:
> We are implementing a project using RabbitMQ.

Cool - and thanks for getting in touch.


> I would be interested in LShift's/CohesiveFT's opinion on the patent
> filed by RedHat, per http://bit.ly/lTFb

That's a very interesting question.  We've brought this to the
attention of the AMQP Working Group, of which we are a member.
Currently the matter is under discussion - so keep an eye on how
things evolve.

The way AMQP works is that ALL members of the WG grant a patent
license to ALL implementers and users of AMQP for patents held that
may affect use of AMQP.  So anything in the specification is
definitely safe from patent abuse and this is upheld by a substantial
body of law and commercial interest.  This is similar to, say, the
W3C.  AMQP is protected from royalty and tenancy abuse, as befits an
Internet protocol.

The problem with this particular patent application is that it covers
something not currently in the AMQP spec.  Were the AMQP spec to
include XML exchanges as an "Extension", then Red Hat would be obliged
to grant licenses to users and implementers, if the patent had been
approved.  We don't expect this particular patent to get approved --
it is too 'obvious'.  In any case we and other members of the Working
Group are taking steps to make sure AMQP stays 100% free and open,
even for extensions of this sort.  We want that because it promotes
interop as the spec evolves, and attracts many more implementers.

We are however very annoyed about the Fear, Uncertainty and Doubt that
actions like this cause.  We are astonished that Red Hat applied for a
patent of this type -- as long ago as 2007 would you believe --
without talking to the Working Group about the best way to protect
AMQP Extensions from patent spam and other abuse such as trolling.

The cheapest and easiest solution would have been to establish
defensive prior art by making this very obvious design very public
indeed, preferably attached to legal language placing the text of the
claim in a safe domain (such as AMQP WG, FSF, or Public).  The fact
that this was not done, raises questions about Red Hat business
practice and casts doubt on their motives.  I suppose for now we shall
all draw our own conclusions.

I hope that Red Hat come clean about their intentions and help to
defuse any FUD.  I do believe they are sincerely committed to the
existing Working Group contracts and would like to see them defend
open standards in the future.  Note that I use 'open standard'
advisedly -- the situation here is very different from the OIN patent
portfolio which is designed to protect Linux.  There is no notion of
interop for Linux, at least nothing like AMQP interop.  The approaches
used by the W3C and IETF are more appropriate here.


> How does this effect RabbitMQ?

Currently not at all.  We shall work to make sure that it never does
and expect to be able to report a suitable guarantee on behalf of the
Working Group as soon as possible.  There are a number of ways to do
this, none of which are terribly difficult, and people in the Working
Group are working as a group to make it happen and do so unambiguously
:-)


> What about the overall effect on the
> adoption of AMQP?

I don't think that scaring users and implementers is good for AMQP
adoption.  This gives Red Hat, and everyone else, a strong incentive
to sort out the mess they have caused.  Then we can all move on,
probably stronger for the experience.

Let me know if this helps,

Alexis Richardson,
RabbitMQ






> Cheers,
> Matt
>
> _______________________________________________
> rabbitmq-discuss mailing list
> rabbitmq-discuss at lists.rabbitmq.com
> http://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss
>