On Mar 8, 2007, at 3:08 PM, Meeraj Kunnumpurath wrote:
Hi,
I have been working on the framework to enable federated
heterogeneous deployment of a logical assembly across one or more
physical runtimes. I was wondering whether we can get an end-to-end
story working that demonstrates the assembly, contribution,
artifact resolution and heterogenoeus federation for the upcoming
TSSS sessions at Vegas and Barcelona.
+1
What I am thinking is having three runtimes, one master and two
slaves, deploy a logical assembly with two components (one Java and
say the other Groovy), each allocated to different slave runtimes
and getting them to talk to each other through a couple of
switchable remote bindings.
So far I have committed the following,
1. Abstractions for physical model which includes physical change
set, component and wire definitions, reference and service
definitions, operation and intereceptor definitions etc.
2. Part of the Java physical model implementation
3. Framework abstractions for marshalling and unmarshalling the
physical model objects
4. Java component definition marshaller implementation
5. New builder framework that understands the physical component model
6. Implementation for the Java physical component builder
7. Federated deployer that integrates the marshallers and the
builders with the component manager and connector
8. Discovery service abstraction for the communication fabric
between the master and slaves
9. JXTA implementation of the discovery service
This is what I think we will have to do to get an end-to-end working,
1. Assembly and allocation on the master
2. Contribution to make resources available on target runtimes
3. Finish the Java container on the slave for the connector
integration
I can sign up to help here
4. Artifact repository
5. An extension container (may be something like Groovy) to
demonstrate heterogeonous federation across different component types.
6. Couple of remote bindings (RMI, AXIS, XFire, Hessian etc) to
demonstrate switchable transport bindings
Happy to help here as well. I've started work on a CXF binding, maybe
we should think about that as well?
7. And anything else I have missed
It is definitely a lot of work for the next two weeks (our session
at Vegas is on the 21st). However, if we divide the work up and
push hard, I am quite positive we can achieve this. I think it will
be an excellent publicity boost for us to have a live application
demonstrating the SCA concepts and heterogeneous federation aspects
of Tuscany, than having just slideware.
Good summary. I'll get stuck into the wiring tomorrow - hopefully
we'll get this all rigged up with the contribution service running
across multiple-VMs.
Jim
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
