[Nml-wg] DCN glossary

Jeroen van der Ham vdham at uva.nl
Tue Jul 8 09:03:01 CDT 2008

Below is a discussion that John and I had that we inadvertently kept off 
the list.

John Vollbrecht wrote:
> On Jul 4, 2008, at 4:37 AM, Jeroen van der Ham wrote:
>> Hi John,
>> John Vollbrecht wrote:
>>> On Jul 1, 2008, at 3:28 PM, Jeroen van der Ham wrote:
>>>> John Vollbrecht wrote:
>>>>> Do you have the latest schema?  I would like to understand how this 
>>>>> all fits together.
>>>>> I have some questions about how multiple circuits can be carried 
>>>>> over a Link.  And how multiple paths can be carried over a sequence 
>>>>> of links.  Perhaps I can ask questions better after looking at the 
>>>>> schema.
>>>> That is a multi-layer description, and that should be handled in the 
>>>> extension of the base schema.
>>> Seems to me this is a single layer description, or a description of 
>>> how a dcn "view"  of a network can be incorporated with a global 
>>> network description.  I am not sure what makes it multi-layer.    Is 
>>> it that building circuits on top of a set of links and nodes?  Sorry 
>>> to be dense, but I don't understand the problem and would like to be 
>>> able to talk about it in a way we both can understand.
>> In my opinion this is multi-layer: you are describing the topology, 
>> and  how several paths are layered over that topology.
>> And to be able to describe what a valid path is, you also need to know 
>> about adaptations, labels, et cetera.
>> So in my view, even if you describe the simple topology, and describe 
>> paths just as a list of nodes, ports and links, then you still need to 
>> know about the technology involved to be able to see if it can all work.
> I think I agree with you.  My confusion is that in other discusssions it 
> seems that there is a single path over a Link - for example there can be 
> only one adaptation per Link implies that there is only one path since 
> there can be many paths and potentially a different adaptation per path.

In NDL an adaptation is a function, and is not tied to a single path. 
The same adaptation can be used many times in the same path. The label 
that is used with the adaptation will be different each time though.

> I agree that one needs to include information about Links, and nodes to 
> create a path.  In particular one needs to know adaptations between link 
> types, bandwidth, and possibly others like VLAN or color.
> I think there is also a difference between describing what is possible 
> to a pathfinder algorithm and what is actually reserved.  What is 
> reserved, including specific resources along the path, becomes a 
> circuit.  A circuit is a path with resources, including time.

I agree that there will be different purposes for network descriptions, 
and that each will have different requirements. However, a base part 
will always be the same. We should be aware of the different 
requirements. But I think that we can now focus on the base schema, and 
later create the extensions.


More information about the nml-wg mailing list