One could argue that wM already had an ESB–Integration Server.
One could also argue (and I think wM has) that Broker and its adapters comprise a so-called “ESB.”
The definition of ESB seems to vary, depending on whom is doing the defining. Here’s one that claims ESB is “an extension of EAI” that adds transformation, “portability” (which they say is the ability to “share data between different computer systems an operating environments”), load balancing/clustering, and failover. I have to ask, how were those “extensions” not a part of “earlier form of middleware” called “EAI”?
This article quotes the Gartner definition of ESB, which in my mind describes many of the products that have been on the market for years. The Gartner definition allows for an ESB to be protocol independent–“Many ESBs also support other communication styles [in addition to web services] that involve guaranteed delivery and publish-and-subscribe; those that don’t soon will.” So, generically, an ESB can be protocol independent–a specific product might not be protocol independent but would still be considered to be an “ESB.”
If we forget about the hype and hooey of all the TLA’s, including EAI, B2B, ESB, SOA, LOL, etc. and the nonsense of whether or not something is a “true” this or that, then we can focus on an integration layer that supports various protocols, formats and such, probably using multiple toolsets and technologies.
Focus on creating an integration architecture and infrastructure that supports what your enterprise needs, using the best-fit tools out there. If one focuses solely on whether or not a tool is an “ESB”, there is a risk of making erroneous assumptions about whether or not the tool has specific functionality that may be needed.
#API-Management#webMethods#soa