Wow, I need to keep up on my specification reading. I missed that one
while reading early versions of the Java to IDL mapping. Did remote
class loading make it into a later version of the spec, or was it
always there?
And it certainly isn't apparent from the RMI-IIOP release and
documentation that this feature is supported. I didn't find
confirmation
of this until I found some mention of it in a FAQ on the IBM RMI-IIOP
web site. It could be better documented.
>From a quick glance through the spec, I couldn't tell whether support
for dynamic stub / valuetype class loading is required. It would be
interesting to know what vendors support this over RMI-IIOP. It would
also be interesting to know what vendors *encourage* the use of this
feature.
-eric
Alan Greenspan wrote:
>
> The RMI-IIOP implementation from Sun dynamically loads classes from the
> server in a fashion identical to that of RMI-JRMP. IMHO, any CORBA ORB
> which claims RMI-IIOP functionality, but can't dynamically load stubs and
> implementations, just simply is not implementing RMI semantics correctly.
> The Sun reference should be used as a barometer.
>
> ORBs compliant with CORBA 2.3 should be able to download stubs and run the
> Sun RMI-IIOP example programs without change. I think this is critical
> functionality for large scale deployment and interoperability of EJB servers
> and EJB clients.
>
> To understand how CORBA implements this functionality, please reference the
> CORBA 2.3 spec. In particular, the pass by value specification,
> http://www.omg.org/cgi-bin/doc?formal/99-10-10 should shed some light on how
> this works.
>
> As for security issues, I think these are well addressed in Java 2 and
> dynamic class loading makes environment configuration much easier and
> extensible than having to deploy clients with all the classes that a server
> can (and might in the future) deploy.
>
> > I thought that RMI-IIOP did *not* dynamically load classes from the
> > server.... this type of thing is not standard CORBA. (Ie, you can't
> > assume what "language" is on the other end... no guarantees that it
> > is Java.)
> >
> > And IMO, dynamic class loading (ie, from the impl server) was one of
> > the biggest drawbacks of RMI-JRMP. It opens up more avenues for
> > security
> > attacks, does not perform as well as local classes, and introduces a
> > constant source of environment misconfiguration.
> >
>
> ===========================================================================
> To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
> of the message "signoff EJB-INTEREST". For general help, send email to
> [EMAIL PROTECTED] and include in the body of the message "help".
===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff EJB-INTEREST". For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".
