qpid-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From William Henry <whe...@redhat.com>
Subject Re: proton: motivation and strategy
Date Fri, 20 Jul 2012 20:03:39 GMT

Wow that was some thread! Sorry I'm not going to respond to any specific part but the overall
themes.

It seems there are two higher level threads:
1. What about proton and where is it's future?
2. What about legacy Qpid and how does proton effect it?

And that is really how it should be.  

The questions we should be asking to figure out the answers to the above are:
Is AMQP more important than "legacy" Qpid?
Is proton about AMQP 1.0 or about "legacy" Qpid? (migrating Qpid to AMQP 1.0)

AMQP and it's broad adoption is the most important. Proton is all about AMQP 1.0 and therefore
it is critical to the expansion and adoption of AMQP.

Legacy Qpid is also very important but in a different way.

Qpid was traditionally about AMQP in terms of the broker. Since AMQP 1.0 the exchange/queue
broker model is less important but still VERY important as a specific AMQP solution for many
users.

(Before anyone gets on my case I will also say that legacy Qpid is also about the Messaging
API and that is also very very important!)

I'm not sure how best these projects get homed in terms of projects in ASF or otherwise but
some thoughts/observations:

Proton, as a library that helps adoption of AMQP 1.0, should not be held back by legacy Qpid.
Proton itself will evolve and is already doing so by the fact that ActiveMQ is already consuming
it. And that's how we want this to happen.  If so called "competitors" (and I don't necessarily
mean ActiveMQ) are consuming proton that's a good thing, and it's also beyond our control.
Proton can be consumed and will be consumed (hopefully) by other projects.  Great! Excellent!
Go AMQP 1.0. Go proton. It's a wonderful thing if proton becomes the de facto library for
advancement of AMQP. 

What about legacy Qpid? What about brokers in general? 

Once "we" put out proton that all became a very different conversation. The tail wagging the
dog.

I see a place where there are AMQP (proton) based solutions. I don't see one broker. I do
see consolidation of effort but I see different solutions. Perhaps some part of the community
continues to invest time and effort in a C/C++ based broker that solves a specific problem.
Perhaps the ActiveMQ based broker solves another class of problems. Perhaps some new "fashionable"
project gets started somewhere else (or at Qpid) that solves a different "routing" problem.
Perhaps someone develops a new language API at Qpid. Maybe someone in Mozilla builds a AMQP
plugin using proton.  Proton has let the cat out of the bag. IF we try to reign it in we might
lose some very good momentum for AMQP. 

What we don't want to happen is that the current Qpid broker "product" or solution should
hold back what proton can potentially do best: drive wider and faster AMQP adoption.  qpid.apache.org
can still be a home for many AMQP based solutions and could presumably still be a home to
proton.  

BUT if qpid.apache.org could/would hold back proton then we have to consider if it is wiser
to somehow separate the efforts in some way (mailing list or otherwise). By the way, that's
little different than it already is today.

So to the folks that make decisions on this list:
Is AMQP more important than "legacy" Qpid?
Is proton about AMQP 1.0 or about "legacy" Qpid?

Can Proton live at Qpid and NOT be held back by legacy Qpid? What's best for Proton and AMQP
in general?

Best regards,
William

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@qpid.apache.org
For additional commands, e-mail: dev-help@qpid.apache.org


Mime
View raw message