Hello,
Below are some of my notes from yesterday's NML Schema teleconference.
In the teleconference we presented the new schema diagram that Martin,
Freek and I made in Amsterdam about three weeks ago. We have made some
significant changes, but mostly based on suggestions at the last OGF
meeting.
One change stands out, the way we define adaptations.
There are now three different kinds of Adaptations:
- Adaptation Function is the way of describing a standard adaptation
(Ethernet to VLAN, OC192 over lambda, etc.)
- Adaptation Service is used to describe the ability to provide an
adaptation function in a node.
- Adaptation Relation is then used to define a specific configured
adaptation.
Freek brought up the point that Links and Crossconnects should also be
relations. The counterargument is that it is then impossible to describe
properties of a Link, such as its capacity.
Freek posed the question what may be missing? He suggested that we
currently do not have Layer and Label objects. We suggested that we
should make examples and apply the schema to use-cases to figure out if
the current schema makes sense. Then we will also quickly find out if we
are really missing things.
Jeff brought up the point that Crossconnects and Adaptations are really
the same kind of things. The difference currently is that adaptations
happen between two logical ports that are part of the same physical
port, and a crossconnect happens between two logical ports that are not
part of the same physical port.
A crossconnect also happens on the same layer, whereas an adaptation
will always happen between different layers.
John inquired about how to group objects to say that they are part of an
IDC. We have the Domain object for that to group different network
objects. The Domain object can then point to an IDC or other
provisioning service.
During the teleconference it became clear that UML is a nice way to get
started, but now that we are getting down to the specific details we
should move to text and examples. This should also allow us to better
see the possible differences (or lack thereof) on Adaptations,
Crossconnects and Links.
Most important points:
- Next teleconference is on July 30th at the same time (16:00 european,
10:00 eastern).
- Martin will submit his work on defining the schema in text before the
next teleconference.
- Freek will prepare some use-cases and examples on cross-connects and
adaptations, so that we can get a sense of how the schema works.
Jeroen.
