"I agree, but it is still a good idea to have a graceful degradation plan available on a separate network if availability is critical. This is made easier by abstracting the core services into libraries/classes that handle these failures in a manner a…" [Mar 12 2008 11:47]
