perfsonar development work

[Loukik Kudarimoti <>]

Maciej Glowiak wrote:
> Roman Lapacz napisał(a):
> > Nicolas Simar wrote:
> > > Definition of pS service
> > > -    Protocol Standard, Syntax (grammar, binary replica of message) 
> > > and Semantics (business logic).
> > > o    Include LS registration,
> >
> > LS registration is a very useful feature but should not be necessary 
> > to treat a service as a part of the perfSONAR. For example there 
> > could be perfsonar infrastructure deployed in a domain without LS.
>
> Yes, but then we don't need LS at all. If we have been developing LS 
> for such a long time we should agree now, that all new services must 
> have capability of registering to the LS. You don't need to use it if 
> you don't want to do it, but if someone takes such a service and wants 
> to run in his environment with the LS, it must support LS registration.
I don't agree with this point Maciej. LS is a very important part of the 
infrastructure but it doesn't mean every service *has* to register with 
it. The advantage of registering with LS is that measurement data 
becomes easily discoverable. This advantage should convince people to 
put that functionality in their services. We should not force them to 
have the functionality just to be called a perfsonar service.

Loukik.
> Maciej