[teiid-designer-dev] [teiid-dev] vdb connector binding names

[Steven Hawkins]

All this is making me like the alternative scenario more.  The only time an explicit source name (aka vdb specific connector binding name) other than the model name comes into place is with multi-source.  And in that case we assume source name == connector binding name.

That gives us updated terminology:

a connector / connetor type - scope deployment wide
a connector binding / connector binding name == source name / jndi name - scope deployment wide
a model name - scope vdb

The additional requirement for multi-source is that they must keep connector binding names consistent across environments.

----- Original Message -----
From: "Ken Johnson" <kejohnso at redhat.com>
To: "Steven Hawkins" <shawkins at redhat.com>
Cc: "Ramesh Reddy" <rareddy at redhat.com>, "teiid-designer-dev" <teiid-designer-dev at lists.jboss.org>, "teiid-dev" <teiid-dev at lists.jboss.org>
Sent: Tuesday, March 2, 2010 10:50:12 AM GMT -06:00 US/Canada Central
Subject: Re: [teiid-dev] vdb connector binding names

I think I agree with this but have few clarifying questions points below.

Steven Hawkins wrote:
> We should be clearer on our terms.  There are:
>
> a connector / connetor type - scope deployment wide
> a connector binding / connector binding name / jndi name - scope deployment wide
> a vdb specific connector binding name (aka source name) - scope vdb
>
> I'm suggesting that in non-multisource scenarios that model name == vdb specific connector binding name.  If there are two models sharing the same connector that's fine.  
>   
Right.  Each model is bound to the same connector but each has it's own 
binding and thus it's own binding name.

The terminology is letting us down here again.  To clarify the models are using the same connector binding.  This is not the same as having the same connector and two connector bindings assigned to two different models.

> At deployment time assigning a connector binding to a model is sufficient for the non-multisource case.  For 
Here, will the vdb-specific-connector-binding-name come into play?  That 
is, will the binding gesture be a mapping of the 
vdb-specific-connector-binding-name to the deployment-wide connector 
binding name?  Or will it be a binding of the model name to the 
deployment wide connector binding name?  I assume the latter but want to 
be sure.

> multisource connector binding association, an additional "source name" attribute would be required.  If the deployer has multiple multi-source models associated to the same set of connector bindings, it would be up to them to ensure that the source name was consistent.  
>   
This could potentially be managed by convention.  That is, when using 
multi-source models, those connector bindings need to be dedicated to 
the multi-source models.  They cannot be shared with standard, 
single-source models or other multi-source models.  It's a restriction 
but should be have minimal impact on the broad user community.

> Alternatively we can infer the source name to be the connector binding name.  This removes the need of having a vdb specific source name (which removes some desirable indirection), but this is how it worked in MMX anyway.
>
> ----- Original Message -----
> From: "Ramesh Reddy" <rareddy at redhat.com>
> To: "Steven Hawkins" <shawkins at redhat.com>
> Cc: "teiid-dev" <teiid-dev at lists.jboss.org>, "teiid-designer-dev" <teiid-designer-dev at lists.jboss.org>
> Sent: Tuesday, March 2, 2010 9:32:49 AM GMT -06:00 US/Canada Central
> Subject: Re: [teiid-dev] vdb connector binding names
>
> Using the model name as the connector name is fine for default scenario,
> unless there are two models that are associated with same connector.
>
> On Tue, 2010-03-02 at 10:11 -0500, Steven Hawkins wrote:
>   
>> Hello all,
>>
>> There seems to be some confusion over what's needed for connector binding names.  To simplify things, there are two basic requirements - we need a known/stable/not environmentally dependent name for each connector binding used in a vdb and there also needs to be a way to associate that name with an actual JNDI resource.  In the case of non-multisource models, the model name itself satisfies the first part.  Only in the multisource scenario do we have an issue with needing an additional vdb specific name for the connector binding.  As a first cut it seems fine to say that both the JNDI mapping and the vdb specific naming of multi-source connector bindings is a deployment time concern.  Then the only design-time concerns become the model name and optionally marking a model multi-source enabled.
>>
>> Any thoughs?
>>
>> Steve
>>
>> _______________________________________________
>> teiid-dev mailing list
>> teiid-dev at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/teiid-dev
>>     
>
> _______________________________________________
> teiid-dev mailing list
> teiid-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/teiid-dev
>