RE: [ASSEMBLY-31] Wiring from a reference with no binding to a service witha binding

[Mike Edwards <mike_edwards@uk.ibm.com>]

Dave,

Comments inline.

Yours,  Mike.

Strategist - Emerging Technologies, SCA & SDO.
Co Chair OASIS SCA Assembly TC.
IBM Hursley Park, Mail Point 146, Winchester, SO21 2JN, Great Britain.
Phone & FAX: +44-1962-818014    Mobile: +44-7802-467431  
Email:  mike_edwards@uk.ibm.com

David Booz <booz@us.ibm.com> wrote on 29/11/2007
13:59:08:

> First to Mike E's proposal with my comments <dab> </dab>:
> 
> - binding.sca is only present on a service or a reference if:
> a) it is explicitly specified
> b) no binding is specified at all (ie then it is the default binding)
> <dab> right </dab>
> 
> - binding.sca is regarded as compatible for wiring purposes with any
> other binding type, and the actual
> binding used is the binding type on the other end of the wire.  If
both
> ends of the wire are binding.sca then
> the runtime is free to choose the actual binding to use.
> <dab> This raises a deployment problem when new clients, which
use
> something
> other than binding.sca, are deployed after a service that exposes
only
> binding.sca.  This would force the service to be re-deployed
in order
> to create the new protocol/binding endpoint. </dab>
> 

In my model, this is not a redeployment.  This
is the creation of a new
endpoint for the existing service component.  I
think binding.sca does
this already due to the ability to use "shortcuts"
when the client and
provider run in the same process (for example).


> binding.sca will be declared as supporting a particular set of intents
> <dab> I assume this is done through the bindingTypes declaration
> so that vendors can decide what intents they want their binding.sca
> implementation
>  to support.</dab>

I hadn't thought of it like that - I thought that
the specifications would
simply lay down a list of supported intents.  However,
I can see the merits
of an approach like the one you describe.  The
downside is that binding.sca
would mean different things on different vendors runtimes.

> if a reference or service uses intents not supported by binding.sca
then
> it is an error if binding.sca is used
> for that service/reference
> <dab> right </dab>
> 
> Second, to Peter's idea:
> In general I agree because it avoids the deployment problem I mentioned
> above and
> because binding.sca should be the 80%, maybe 90% use case for component
> interactions.  The additional burden for the 10-20% is not onerous
IMHO.
> 
> 
> Going back to the original problem, another alternative might be to
have
> slightly different rules for consumers versus providers.  I don't
like
> asymmetry, but
> let's try this out anyway for a moment.  What if we allowed a
reference
> with only
> binding.sca to wire to any provider binding, AND we require that any
> non-binding.sca reference to match (binding type match) a provider
binding?
> 
> 1) This avoids the deployment problem.
> 2) This would allow a reference binding.sca implementation to morph
itself
> into
> any provider binding (ed: which is one of the reaons that SCA should
> _never_
> try to create a cross vendor interoperable binding.sca). It enables
easier
> access
> to the one of the 10-20% use cases (domain component using another
> Domain component that is also exposed outside the domain).
> 3) It would also work with callbacks as the provider only has to be
able to
> speak
> what he's already configured to speak.

This is a subset of my proposal.  Depends on
whether you think that a
provider using "binding.sca" is capable
of running multiple endpoints
simultaneously for different protocols.  I think
that this is OK, so I don't
need to subset the proposal and it remains symmetric.

Interesting to see what other folk think on this point.


> 
> 
> 
> Dave Booz
> STSM, SCA and WebSphere Architecture
> Co-Chair OASIS SCA-Policy TC
> "Distributed objects first, then world hunger"
> Poughkeepsie, NY (845)-435-6093  or  8-295-6093
> e-mail:booz@us.ibm.com
> http://washome.austin.ibm.com/xwiki/bin/view/SCA2Team/WebHome

---

Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 741598.
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU