Re: [soa-rm] On the topic of reference models...(now ESB)

["Jeffrey A Estefan" <Jeffrey.A.Estefan@jpl.nasa.gov>]

Hi Joe,

 

I've read the Sonic white paper (and several of their white papers in the past).  They [Sonic] will continue to insist that they invented the first ESB, or at least were the first to coin the term; and so will Fiorano, and IONA, etc., etc.  I attended the last several Enterprise Architect Summit conferences and Chappell and van Huizen from Sonic always claim they invented the ESB; they make a point of it.  And then you'd actually get the Fiorano CTO up on the same stage we these guys and claim that they [Fiorano] created the first ESB.  My point is to approach these guys with caution and try to remain steadfast in seeing through the marketing hype.

 

I don't think we will want to include a definition for ESB in our SOA-RM because although ESB is not a product, it is an architectural best practice for implementing a SOA and thus better suited for an architectural blueprints or reference architecture effort.  Incidentally, implementing an ESB requires an integrated set of middleware services (i.e., technology) so again, out of our scope.  It is important to note, however, that the ESB concept is intended to support multiple architecture styles, e.g., SOAs, message-driven architectures, and event-driven architectures.

 

So in summary, I would not "sign-up" for any recommended "RM" or "standard lexicon" or whatever a specific middleware vendor wishes to push, including Sonic.  They have a fine offering in this space but again, I think our job (and I know you agree) is to remain vendor-neutral.  Frankly, after reading their white papers, they are much more focused on ESB architecture than a creating a true RM.  You can see that they even refer to the W3C's WS Arch WG definition of SOA rather than creating their own definition for SOA.

 

V/R, Jeff E.