Disagree, Gregg. An effectiveness of the system is not an absolute value, it has a baseline - requirements and environment. As a consumer, I do not care how effective a system was being accessed via RMI/IIOP if it responds slow via HTTP, do I?
Some systems cannot work in new environment, e.g., in Web, because they were built for a few users only ( single-threaded). My point is that placing a Web Service or any other interface on such system does not make it a service automatically. The wrapped app has to proof it can work effectively in new conditions first. This risk I am talking about. - Michael ----- Original Message ---- From: Gregg Wonderly <[EMAIL PROTECTED]> To: [email protected] Sent: Wednesday, July 2, 2008 5:03:24 PM Subject: Re: [service-orientated-architecture] Legacy into SOA (was Vandersluis on a Data Abstraction Layer's Benefits) Michael Poulin wrote: > Congratulation, Patric! > > Patric has demonstrated that a non-service oriented legacy system ( ERP > in his example) is an obvious risk, if not a threat, to the SO based > solution. What we gonna do about it? > I am confident that just wrapping legacy systems with Web Services and > pushing a process via an ESB is a risky business. If one takes Sun's > Composite Application model for an analysis, how reliable the entire > system is? What we have to do to mitigate this risk either in services, > or in the infrastructure , or in both? I.e.; A system that has a particular style of interaction doesn't change in its effectiveness if you just change how the interaction is transported across the network.... Gregg Wonderly
