This is technically feasible (see other discussion on this thread) but I think we need to be careful about how wide open we are making the services within a Tuscany SCA domain. I don't think we want every service within every domain to be wide open for anyone to reach down from the sky and call it. As the owner of a service, I would want to decide whether it is exposed for local calling, distributed calling within my domain (which can consist of multiple runtime nodes), or revealed to other domains via some global locateService API. I think the right way for me to exercise this control is to create a self-reference or a service reference within my domain, then selectively pass this out to other domains in a manner that I can control, e.g., in a service registry that is able to implement a governance policy.
Simon Kevin Williams wrote:
I am interested in extending the current locate service capability to support the location of services that may be running within the same domain but a separate jvm. The most straightforward approach may be to dynamically create and return a service reference using the same Tuscany runtime elements that now create service references for injection in support of the @reference annotation. Any thoughts or pointers would be appreciated. Thanks, --Kevin --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
