Re: [Nml-wg] Topology Model and Pathfinding
On 10/02/2010 15:01, John Vollbrecht wrote:
If I don't have a link between networks, how do I describe the topology? My suggestion has been that we describe the topology as being networks interconnected at points (perhaps that is a bad name, but it is what G.805 uses). Then it looks like your picture except that instead tying the networks together with a link I tie them with a point.
G.800 and G.805 also have links. Even if a stand alone link is a network, why can't we treat it as a special case and leave that in as a way to describe the inter-connection point between two networks? Basically two edge-points are requird to describe the interaction between two neighboring networks. I propose to use the ports at either end of a Link, which fits with our current network model. So far I have still not seen a convincing argument why this does not work, and why we would need to introduce additional terminology to describe the interaction point. Jeroen.
On Feb 10, 2010, at 6:14 PM, Jeroen van der Ham wrote:
On 10/02/2010 15:01, John Vollbrecht wrote:
If I don't have a link between networks, how do I describe the topology? My suggestion has been that we describe the topology as being networks interconnected at points (perhaps that is a bad name, but it is what G.805 uses). Then it looks like your picture except that instead tying the networks together with a link I tie them with a point.
G.800 and G.805 also have links.
Even if a stand alone link is a network, why can't we treat it as a special case and leave that in as a way to describe the inter-connection point between two networks? well, it isn't really a special case as far as I can see. the picture I added shows that.
Basically two edge-points are requird to describe the interaction between two neighboring networks. I propose to use the ports at either end of a Link, which fits with our current network model. yes - you could say there are ports on either side connected at a point. ou So far I have still not seen a convincing argument why this does not work, and why we would need to introduce additional terminology to describe the interaction point. I am sorry -- but I don't see how your solution works. Could you show a topology of connected networks that doesn't include links? It seems to me that you are trying to force fit something that is slightly off.
I am not sure how to resolve this without exploring possibilities. Do you think that my suggestion for how to create a graph with networks as vertices and points as edges will not work? If it will work is there a way to fit it into the NML model or would the model have to change to accommodate it? John
On 10/02/2010 16:41, John Vollbrecht wrote:
I am sorry -- but I don't see how your solution works. Could you show a topology of connected networks that doesn't include links? It seems to me that you are trying to force fit something that is slightly off.
What is wrong with Links that we can't include them in the inter-domain topology? Jeroen.
On Feb 10, 2010, at 7:59 PM, Jeroen van der Ham wrote:
On 10/02/2010 16:41, John Vollbrecht wrote:
I am sorry -- but I don't see how your solution works. Could you show a topology of connected networks that doesn't include links? It seems to me that you are trying to force fit something that is slightly off.
What is wrong with Links that we can't include them in the inter-domain topology? They don't exist by themselves in the model. They are part of a network, just like nodes are part of a network.
Jeroen.
participants (2)
-
Jeroen van der Ham -
John Vollbrecht