perfsonar development work

[Loukik Kudarimoti <>]

Jeff W. Boote wrote:
> Joe Metzger wrote:
> > Hi,
> > My management has asked if perfSONAR could support some of the
> > visualizations that DREN is starting to use.
> >
> > http://www.wcisd.hpc.mil/~phil/Planet-DREN/
> >
> > This is an interesting integration of Toplogy, Utilization,
> > Latency and Bandwidth measurements.
> >
> > Is anybody working on this?
>
> (This is Phil Dykstra's stuff, right? Matt and I were helping him get 
> some OWAMP data collected for the Abilene/DREN peering points... But, 
> I don't know if he is integrating any of that into these 
> visualizations. This is very nice work.)
>
> I guess it depends on what you mean by 'supporting' those 
> visualizations... I'm sure there are similar things in the works in 
> CNM and Nemo (and perhaps Vedrin has ambitious plans for perfSONARUI).
>
> I suspect the easiest way to support these types of visualizations is 
> to make integrating the perfSONAR data into other applications as easy 
> as possible. For example, if Phil could have his application consume 
> the data that is made available from various perfSONAR MA's.
>
> The trick would be to make sure we have geographical references for 
> all subjects in the MA's.
>
> I don't think anyone could adequately answer how difficult it would be 
> to integrate pS data into Phil's application without knowing what kind 
> of data collection his stuff uses. (And Phil probably doesn't know 
> enough about perfSONAR to know from his side.)
Looking at this application, the first question that comes to my mind 
is: Is it possible for an application of this kind to retrieve all the 
necessary data via NMWG interfaces so that it can come up with such 
visualizations? Knowing a bit about Google Earth and similar 
applications, I think we can say "No, It is currently not possible". 
This is because we do not have a topology service and also the 
measurement data that we provide does not contain any geographical 
information required by such applications.

Supposing we were ready i.e., we have the geographical information and 
we have the topology service, my next question is "Who is going to 
develop such applications?". I think that sometime in future we need to 
come up with a strategy which will encourage developers (apart from the 
current visualisation tool developers) to work on a variety of 
visualisation tools.

My opinion is similar to Joe's. We are probably looking at 6-12 more 
months of effort on services and other issues. After this, I hope that 
we will be in a position to think about some strategies for getting more 
visualisation tools developed by independent developers.

Loukik.


> But, I think these kinds of integrations are the way forward. Similar 
> to the way that we have integrated into existing RRD infrastructures 
> to make data available, we should integrate into existing 
> visualization and analysis tools.
>
> jeff