Jason Zurawski wrote:
I believe this makes the most sense as a service, I will add this comment into the ticket as well.
Yeah. It felt a bit awkward to describe a fixed adaptation as a service. Mostly because I (mis?)interpreted "service" as "dynamic service". A SFP contains an Ethernet-over-a-wavelength adaptation. You may say that the wavelength provides an Ethernet service, even there is nothing dynamic about it (the Ethernet may in turn provides dynamic services, but that's a different Service description). Three reason to describe it as a service: 1. That's what we came up with at OGF 33. 2. I want to pick one, not allow both. 3. NML will mostly describe the dynamic adaptation, only rarely static adaptations. (3) (3) Path finding uses dynamic adaptations; One often monitors dynamic services, etc. Static adaptation are mostly used within a domain and that info is -I think- rarely exchanged with other domains. Freek PS: Sorry for the many mails today. I've recently been busy applying NML to a real-world monitoring scenario, and this is more or less the culmination of issues that I've found. On the bright side, I was happy to see how complete NML is by now. The only downside was that I ended up writing a 2000-line XML file by hand. According to my wife I was mumbling about "closing tags" in my sleep ;).