What is a best of breed ESB (four years on)?

Four years ago my colleague, Steve Craggs, wrote a seminal paper for the Integration Consortium called "Best of Breed ESB" (available from the Integration Consortium web-page and some other places such as here).  It was seminal in that it put an early stake in the ground defining what to look for in an ESB.  At the time, the concept of an ESB was pretty new and the definition was very vague (usually a single sentence from Gartner).  Four years on the ESB category has matured - the major vendors now all have something they call an ESB - but there is stil confusion to dispell and diversity in product offerings which we will be covering in an upcoming new edition of this paper.

Back in 2002, Steve's paper set the ground rules for what to look for in an ESB by setting out and discussing what users should expect to see:

"Fundamental ESB characteristics:
• XML, messaging, transformation, intelligent routing services
• Basic connectivity (Web Services, J2EE Connectors, JMS)
• Service-oriented architecture
• Support for highly distributed deployments
• Manageability
Key, value-add characteristics:
• Robustness
• Scalability and Performance
• Security
• Breadth of connectivity
• Development / Deployment toolset"

Unlike many of the analyst papers that followed it did not simply provide a lengthy ticklist of standards to be adhered to - rather it discussed why capabilities are needed and what alternative approaches were taken by vendors to provide those capabilities.  This was particularly important as not only did it clear confusion and FUD that was already gathering in the space, it also reflected the reality that the diversity of use cases and organisation types considering ESBs meant that there could be no single magic ticklist.  Back in 2003 when the paper appeared, I was CEO of PolarLake, one of the first ESB vendor, and it resonated because it solved a major problem for vendors and buyers alike: It provided a framework for matching requirements against products and comparing products. 

Four years on looking at the vendor offerings, while the basics are now settled there is even more diversity at a value-add feature level as well as new market dynamics (OSS has become a major force in ESBs).  On top of which, great diversity remains in use cases.  For these reasons, now seems like a good time to update and revise Steve's earlier work and I am now in the process of producing a new Best of Breed ESB paper.  While taking a simiar approach and covering some of the same bases, there will be a very different focus reflecting the major issues being faced by users today.  In particular: how do ESBs help with the SOA skills shortage and how do they support service reuse - as well as a whole raft of deployment issues.  The paper should be out in September as I am now in the process of being briefed by vendors.    

Watch this space...

Ronan

Comments

[Kristen Puckett]

With an emphasis on ease of use and rapid deployment of vendor-independent components, Bostech's ChainBuilder ESB includes a set of Eclipse plug-ins to simplify the creation and customization of components for a JBI-compliant ESB.

A java developer can leverage the run-time advantages of JBI components but concentrate just on creating the core logic, since ChainBuilder ESB generates the underlying JBI-compliant code automatically.

ChainBuilder ESB 1.1 is available for immediate open source download at: http://www.chainforge.net.

To include Bostech in the vendor analysis, you can contact Sales and Marketing at contactbostech@chainforge.net.

[Steve Craggs]

Thanks for that subtle piece of advertising Kristen! I am sure we will include you as we carry out our research for the new paper.

Steve