perfsonar development work

[Jason Zurawski <>]

Ulisses;

> > Ulisses, I don't quite notice significant changes in the above xml in 
> > comparison to the ones you sent in the last email. I am not sure if one 
> > data element can be put inside another. I see this being done in the xml 
> > you mailed me as well. Has Jason confirmed this to you? Maybe its best 
> > to wait for Jason to come back.
> >
> >
> > A standard message is usually of this kind. But again, I am only 
> > guessing what you are trying to achieve so I might not be right here.
> >
> > <nmwg:message id="msg1" type="TOPSDownloadDBResponse" 
> >                        xmlns:nmwg="http://ggf.org/ns/nmwg/base/2.0/";
> >                        
> > xmlns:nmwgtopo3="http://ggf.org/ns/nmwg/topology/l3/3.0/";>
> >
> >    <nmwg:metadata id="TOPSDownloadDBResponseMetadata">
> >        <nmwg:subject id="sub1">
> >            <nmwgtopo3:link id="link1">
> >                <nmwgtopo3:name type="logical">test</nmwgtopo3:name>
> >            </nmwgtopo3:link>
> >        </nmwg:subject> 
> >    </nmwg:metadata>
> >
> >    <nmwg:data id="TOPSDownloadDBResponse" 
> >    metadataIdRef="TOPSDownloadDBResponseMetadata">
> >    </nmwg:data>
> >
> > </nmwg:message>
> >
> >     
>
> is extrange to see the data (links in this case) in the metadata element.
>
>
>   


The topology elements were originally designed to to be the 'subjects' 
of the various measurements, and to keep with that spirit (unless a 
compelling reason can be found as to not perform things in that same 
manner) they should be kept that way.  Even though they are the 'data' 
of your response, they are still metadata in the system.  The LS does 
something similar where the 'data' of a request is returned in 
'metadata' elements, so the idea is not too far fetched. 


> > No. I meant to ask what are the possible types of requests and responses 
> > that your Topology service will be able to handle? Have you discussed 
> > these requests and responses with Jason and Martin?
> >     
>
> no
>
>   

We should do this soon, either through call or mail.  Look to the LS/MA 
examples in svn or if anything was written in the previous Topo service 
documents to help you make the necessary requests and responses.


> > I mean to ask whether the requests and responses that you have discussed 
> > and agreed to use are the same as the ones your service is generating.
> >     
>
> nobody never told me that I had to conform to other than the *.rnc files
>
> I supossed I was the one to define the document, and that I only had to 
> conform to this .rnc files.
>
>   

The rnc documents are just syntax, they really don't explain the 
semantics of how a service will perform.  Even so the rnc documents 
currently in use just describe the topological elements at the layers, 
not really too much on how your service will use them.

-jason