perfsonar development work

["Ivan K. Koga" <>]

The name of the source or destinations are just internal descritive 
parameters. They are not imposed to be common, but it is reasonably that 
it is the same if the host is the same. (Ex: one machine is named 
amp-unifacs in a amp measurement point and could be named amp-unifacs2 
in another measurement point).

Jeff, Do you suggest to make change in this:

   <amp:parameters id="param1">
     <nmwg:parameter name="source">amp-unifacs</nmwg:parameter>
     <nmwg:parameter name="destination">amp-ufsc</nmwg:parameter>       
  
   </amp:parameters>

To this?:

   <amp:parameters id="param1">
     <nmwg:parameter name="src.commonname ">amp-unifacs</nmwg:parameter>
     <nmwg:parameter name="dst.commonname">amp-ufsc</nmwg:parameter>   
      
   </amp:parameters>

I think this change would describe better the parameters. I agree with 
Jeff.

Ivan



Jeff W. Boote escreveu:
> Jason Zurawski wrote:
> > Ivan;
> >
> > > I hope this explains better how amp measurements work.  :-)
> > Yes, very helpful.  Attached is my next pass (Martin/Jeff, if you 
> > have time could I get some input?)
>
> This looks fairly reasonable to me. The only question I might have is 
> about the source/destination parameters. If those are intended to be 
> ways of adding 'common name' tags, I would prefer a more descriptive 
> parameter name. And, something that would allow other contextual 
> information about the dst/src to be specified as well.
>
> For example:
>
> src.commonname
> dst.commonname
>
> Then, if we wanted to add other contextual information about the 
> endpoints it is clear how to do that using the parameters.
>
> (Alternatively, endPoint could be extended to add these as direct 
> attributes or elements. Or, we could even add a parameter bag in to 
> endPoint... I don't think it really matters which way we go - as long 
> as it is clear how to add additional contextual information about each 
> endpoint.)
>
> jeff