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.
Jeroen.
