perfsonar development work

[Jason Zurawski <>]

All;

> > > We are discussing what information services should register into LS, 
> > > besides service URL.  Recently, Jason advised to recognize service 
> > > type by eventtypes, rather than by arbitrary strings like one found 
> > > in "servicetype" parameter.
> > >
> > > This seems to work fine with RRD MA services, but not all other 
> > > services submit information about eventtypes when registering to LS. 
> > > Is there a guidance what schema should be supported in this case?
> >
> > Could you be more specific? Which services are not registering 
> > eventTypes? This disrupts the dLS (and gLS) summarization model as 
> > well...
> So far I am aware of flowsa MA, developed at Surfnet ,  and  BWCTL and 
> Hades by Win-Labor.  Verena is going to add support for eventtypes, 
> but was asking of the syntax to be supported. Could you please advise 
> her?


I am not sure I understand the request, services should be registering 
with the LS in a manner similar to the RRD MA etc. where they register a 
Service metadata description, and the metadata elements they are 
personally responsible for.  These metadata elements *should* include 
eventType elements for the particular data they pertain to.  The 
examples that can be examined are related to the RRD MA (this is the 
only one I personally know of, perhaps there are more).


> > I think we should concentrate effort to fix this on the service 
> > registration side, not by adding heuristics on the client side (which 
> > would then need to be copied to every other client).
>
> Yes, sure - we've already introduced some perfsonarUI code that relies 
> on eventtypes registered in LS and are trying to figure out how it 
> will work with services other than RRD MA. 


I can't really stress this enough: discovery will not work unless 
services are registering in the same manner.  If a service does not make 
available to the LS:

a) Address information (ipv4 or perhaps ipv6)
d) Domain information (fully qual hostname will do)
c) any and all possible eventTypes

Discovery will be meaningless.  If service authors have questions or 
concerns about this, lets start talking about it very soon, and if you 
haven't done so *please* read the LS API document, this should explain 
everything:

http://anonsvn.internet2.edu/svn/nmwg/trunk/nmwg/doc/dLS/gLS/phase_1_color.html

-jason