[Nsi-wg] additional call notes

[John Vollbrecht]

A couple additional notes from the call this morning

We had a lot of discussion about parameters and the difference between  
abstract and technology specific parameters.
We had a list of parameters in the issues doc, and Guy will edit these  
into the strawman doc.

Discussion about the parameters
	- abstract parameters seem to imply a range - e.g. a 100mb depends on  
what exactly what technology and network are used, whereas a 10GE  
interface is specific about what is provided at least at the interface

	- Tomohiro made the distinction between quality parameters for the  
connection as a whole and parameters at the edge point.  Edge point  
parameters are things like VLAN or slot numbers, quality parameters  
are things like jitter or drops.  Edge parameters are required, some  
quality parameters are optional (some may be required).
---

We talked about whether Waves and fiber bundles were alternative  
connection topologies.  We had discussion about whether waves were  
appropriate as raw input or if waves always started as bits.  If waves  
start as bits, then it seems to me there is an adaptation to carry  
them on waves, but the underlying "type" of connection is bits.

Joe pointed out that waves were being switched in labs and that  
standards should handle the future where possible.  Also fiber bundles  
are less problematic and can be switched, and their characteristics  
are not bits.

So a question I have is whether NSI should be interested in data  
(bits) connections or also in generic connections like waves or fiber  
bundles.  This is to be determined - I would be interested in comments.

There is also a need to describe L3 and perhaps L4 connections if we  
want to do that.  It might be possible to say these are possible but  
concentrate on L2 for first pass of the document.
---

A discussion about different services that are used by and perhaps  
provide by a NS included the suggestion by Chin (which seems right to  
me) that Topology sharing be described basically for inter NS but  
allow it to be used by others.  I think the intent is that we should  
include this as needed to support inter NS capability and assume that  
will be sufficient for other users.

--\
Please comment or add to any of these -

John