1) Always proxy inbound requests between the external representation and your consumption 2) Always Facade everything external
I actually wrote a transport agnostic framework (with dynamic interfaces) back in 2000 which eventually supported WS/EJB/RMI and local. One slight problem with the view below though is that it assumes that one thing replaces another, it doesn't it builds on it either technically or by considering the old stuff legacy. CICS/COBOL is over, but that doesn't mean there aren't billions of dollars going over CICS transactions. The 2nd point comes back to Peter Deutch's 7 fallacies of networking computing and something I blogged about (http://service-architecture.blogspot.com/2006/10/axis-2-maven-and-problems-with.html) a while back. Failure is the "norm" in distributed systems and things like graceful degredation will have to become more important. On 19/12/06, Alexander Johannesen <[EMAIL PROTECTED]> wrote: > > > > > > > The end of SOAP : > "It won't happen at once, it wont be overnight, but one day SOAP will > be over. We will look back and wonder "what were we thinking". It will > be up there with ActiveX, EJB2, and other things that we will describe > as mistakes that should never have made it past the powerpoint stage." > > "Incidentally, this shows a problem with relying on any external SOAP > or REST service for some mission-critical role in your own code. How > can you be sure that one day your service provider won't turn it off?" > > http://www.1060.org/blogxter/entry?publicid=4B4C0B46DC743154ECB68300531D6A04 > > No one can solve the problem of people turning their services > completely off, but how do we make sure that switching from one to > another is as painless as possible? > > Alex > -- > "Ultimately, all things are known because you want to believe you know." > - Frank Herbert > __ http://shelter.nu/ __________________________________________________ >
