perfsonar development work

[Roman Lapacz <>]

Jason Zurawski wrote:
> All;

Hi
> > > I'm probably going to get yelled at because what I'm going to 
> > > suggest is more complex - but I really don't like the timeType 
> > > parameter solution. In my opinion, all parameters that are of 'type' 
> > > time should be set using an nmtm:time element.
> >
> > After talking with Jason, I decided I should have put an example in 
> > here to show what I'm talking about.
> >
> > So, for the example Roman sent, the response would contain:
> >
> >  <nmwg:metadata id="meta2select">
> >    <select:subject id="iusub2" metadataIdRef="meta1"/>
> >    <select:parameters id="param1">
> >      <! string start time will be interpreted as UNIX timestamp !>
> >      <nmwg:parameter name="startTime">1121472000</nmwg:parameter>
> >      <nmwg:parameter name="endTime">
> >        <nmtm:time type="UNIX" value="1121904000"/>
> >      </nmwg:parameter>
> >      <nmwg:parameter 
> > name="consolidationFunction">AVERAGE</nmwg:parameter>
> >      <nmwg:parameter name="resolution">60</nmwg:parameter>
> >    </select:parameters>
> >    
> > <nmwg:eventType>http://ggf.org/ns/nmwg/ops/select/2.0</nmwg:eventType>
> >  </nmwg:metadata>
> >
> > 1. The timeType parameter would go away completely.
> > 2. Any parameter that is used to specify a time value could be 
> > specified using a string (like now) but it would have to be a unix 
> > timestamp in ascii string format.
> > 3. Any parameter that is used to specify a time value would need to 
> > accept a nmtm:time element as the value.
> > 4. There would be no way for a client to specify what timeType they 
> > want. We should create a time element API for creating time elements 
> > represented with different types.
>
> So the issue still remains of what to do for 'display' of values (not 
> necessarily the same as the time that will be used in the select 
> statement).  I still believe that this is best in a separate parameter 
> block, and as Roman pointed out this couldn't be in the same md block 
> currently.  I would propose some sort of additional filter chaining as 
> well (although I still think we need to abuse the notion of timeType a 
> little further, since this is more of a presentation element than a 
> time value):
>
> <nmwg:metadata id="meta2select">
>   <select:subject id="iusub2" metadataIdRef="meta1"/>
>   <select:parameters id="param1">
>     <! string start time will be interpreted as UNIX timestamp !>
>     <nmwg:parameter name="startTime">1121472000</nmwg:parameter>
>     <nmwg:parameter name="endTime">
>       <nmtm:time type="UNIX" value="1121904000"/>
>     </nmwg:parameter>
>     <nmwg:parameter name="consolidationFunction">AVERAGE</nmwg:parameter>
>     <nmwg:parameter name="resolution">60</nmwg:parameter>
>   </select:parameters>
>   <nmwg:eventType>http://ggf.org/ns/nmwg/ops/select/2.0</nmwg:eventType>
> </nmwg:metadata>
>
> <nmwg:metadata id="meta3" metadataIdRef="meta2select">
>   <nmwg:parameters id="param1">
>     <nmwg:parameter name="timeType">ISO</nmwg:parameter>
>   </nmwg:parameters>
>   <nmwg:eventType>http://ggf.org/ns/nmwg/base/2.0</nmwg:eventType>
> </nmwg:metadata>
>
> <nmwg:data id="data1" metadataIdRef="meta3" />

I'd like to propose a small change. Take a look:


<nmwg:metadata id="meta1">
 <netutil:subject id="iusub1">
   <nmwgt:interface>
     <nmwgt:ifAddress type="ipv4">100.1.2.3</nmwgt:ifAddress>
     <nmwgt:direction>in</nmwgt:direction>
   </nmwgt:interface>
 </netutil:subject>
 
<nmwg:eventType>http://ggf.org/ns/nmwg/characteristic/utilization/2.0</nmwg:eventType>
</nmwg:metadata>

<nmwg:metadata id="meta2select">
 <select:subject id="iusub2" metadataIdRef="meta1"/>
 <select:parameters id="param1">
   <! string start time will be interpreted as UNIX timestamp !>
   <nmwg:parameter name="startTime">1121472000</nmwg:parameter>
   <nmwg:parameter name="endTime">
     <nmtm:time type="UNIX" value="1121904000"/>
   </nmwg:parameter>
   <nmwg:parameter name="consolidationFunction">AVERAGE</nmwg:parameter>
   <nmwg:parameter name="resolution">60</nmwg:parameter>
 </select:parameters>
 <nmwg:eventType>http://ggf.org/ns/nmwg/ops/select/2.0</nmwg:eventType>
</nmwg:metadata>

<nmwg:metadata id="meta3presentation">
 <presentation:subject id="iusub3" metadataIdRef="meta2select"/>
 <presentation:parameters id="param1">
   <nmwg:parameter name="timeType">ISO</nmwg:parameter>
 </presentation:parameters>
 
<nmwg:eventType>http://ggf.org/ns/nmwg/ops/presentation/2.0</nmwg:eventType> 

</nmwg:metadata>

<nmwg:data id="data1" metadataIdRef="meta3presentation" />



- I replaced simple base chaining (metadataIdRef as attribute of 
metadata element) with chaining via subject element (my understanding of 
the base chaining is the solution of dividing metadata element in order 
to reuse its parts)
- I introduced new namespace 'presentation' in order to indicate the 
purpose of the parameter timeType inside

What do you think?

Roman