AMQP UK, Thursday, 4th August 2011
The AMQP protocol’s broad adoption relies upon the Magic Triangle.
Earlier this month I had a chat with John O’Hara, the founder of the AMQP Working Group and the Global Head of Architecture for Commodities at Bank of America. We covered many topics; however the overriding theme was regarding the future of AMQP and the MQ market in general.
We both agreed that the cloud MQ market will most likely follow the trajectory of the web hosting market. 20 years ago you would have had to buy a server and install IIS, Zeus or some other expensive application to host your website. Doing so gave you a hugely configurable system that was often prohibitive in cost. Since then the market has commoditised significantly and now you can signup online and get a rather feature rich service for a few dollars a month. While cheap web hosting does not work for everybody, its availability has increased demand for websites, lowered prices and grown an industry of ‘mom and pop’ hosting companies….many of which have been swallowed up in market consolidation.
But, before the cloud MQ market can take off we need a more broad adoption of AMQP by the general developer community. This needs native AMQP clients installed by default and widely supported by OS and hardware developers, put simply, you would need the full participation of the main operating system stack suppliers such as Microsoft, RedHat and VMware. John described these companies as forming a ‘magic triangle’ of influence and luckily enough, each of these companies is committed to releasing an AMQP 1‧0 compatible service or product. In doing so they will most probably release native clients and thus paving the way for broad adoption. It is the growth of pervasive AMQP clients and a strong and healthy developer community that companies like stormmq will rely upon.
So, why should companies such as stormmq, INETCO, Microsoft, RedHat and VMware develop AMQP based services and products? Put simply, where is the money? We believe that customers will need some of the following services and products:–
- Stand alone brokers
- Such as Apache Qpid
- Or VMware’s RabbitMQ
- Messaging Appliances
- Wrapping it in sheet steel is a proven market
- This is what Solace do
- And Cisco could do
- Management utilities to bring it all together
- eg The Candle model
- These allow people to move from legacy products like TIBCO
- Or in‐house solutions
- Business Transaction Monitoring
- The INETCO model
- Bastion technology
- Routers that bridge middleware
- Specialised deep‐packet inspecting AMQP firewalls
- Cloud based services
- stormmq are the only supplier in this market at the moment
- There are other suppliers getting ready to release
- Embedded devices
- The ‘internet of things’ is going to need integration into the enterprise
- M2M will require middleware that can manage the large flows of data
There are many more potential use cases for the AMQP protocol, but getting broad adoption of the protocol relies upon the Magic Triangle. We are in the early days of the commoditisation of AMQP, and as history has shown us, the only way is up from here.