I first drew this chart back around 2004 for my friend Alexis Richardson. At the time I referred to it in the context of a proprietary research methodology, but I don’t want trademark lawyers chasing me – hence the thesaurised title for this post.

The point was very simple – we had standards based protocols for everything except the most useful case – asynchronous and reliable. There were of course protocols that occupied that corner, but they were proprietary – fine if you’re doing something within an organisation and willing to pay for it, but not so good if you’re trying to get services to work across organisational boundaries.

At the time many people were trying to bend HTTP out of shape so that it could be asynchronous and reliable with efforts like WS-RM, but I felt that what the world really needed was a simple standards based protocol for message oriented middleware (MOM).

Of course my prayers were answered not soon after, with the announcement by my friend John Davies at the Web Services on Wall Street conference[1] of the impending release of AMQP, and Alexis went on to found the hugely popular and successful RabbitMQ. The brave new world was going to look like this:

So why am I writing about this 7 years after first scribbling on a whiteboard? Well… it seems that some people still don’t get the joke. I found myself drawing the same chart again just this week whilst in the midst of a discussion with a SaaS provider on notification mechanisms. The unpleasant surprise was that people found it useful and informative (rather than obvious and patronising).

I’m going along to PubSub Huddle later this week, where I’m sure I’ll find loads of people doing cool stuff with AMQP. I just hope that I don’t have to wait too much longer for this all to become mainstream. For some time I’ve shared the vision of John O’Hara, and others in the working group, of AMQP becoming the ubiquitous interconnect[2]. Hopefully the time for that vision to become reality is upon us now that AMQP 1.0 is out the door.

[1] I recall that John didn’t have approval to talk in public about what was at the time time an internal project, but he went ahead anyway.
[2] John did a great piece in ACM ‘s Queue magazine on this (warning PDF and some scrolling required).