Hi Radak- This helps. I was just curious - just trying to understand the structure. I will be working on one of the Challenge topologies today and I'll post it when it looks like it is suitable to be used. Jerry On 7/27/11 5:03 AM, Radek Krzywania wrote:
Hi Jerry,
I’ve followed the guide to build the topo. Maybe I misunderstand something, but this I show I feel about the topology. So:
-- Why do you have a 1G switch matrix defined with no ports? and why do you have a Node with two Switchmatrix objects? – Indeed there is no connection to 1G switch. This should be removed. The purpose is that some of the ports are 1G and some are 10G interfaces. The first ones do not support VLAN translation, the second ones do. There are some 1G interfaces in the real topology at this moment, but they are under migrating process to 10G. As we do not include data plane at this moment, it doesn’t really matter. Thus I’ve set all ports in topology to 10G capacity. I will remover a 1G switchmatrix, and bring it back only if needed just for clearance
-- Why have you defined hasPort relationships from both the Node object and the Switchmatrix object? This seems redundant...? Would the proper structure have the Node->Switchmatrx->Port(s) ?? – according to the guide sent by Jeroen = “(…) SwitchingMatrix class to describe the internal functioning of a Node. We use the hasPort and canSwap relations to describe between which Ports it can switch connections, and whether the labels can be swapped while switching.” The ports are attached then to both Node and SwitchMatrix. The case is that multiple SwitchMatrix can be assigned to single Node, indicating which ports can be switched. E.g. if you will have 4 ports in single node, but you can switch only between 1 & 2, and 3 & 4, two switchmatrix will defined for proper ports pairs. Please mind that in our case a Node is a network, thus switching between all ports may not always be possible.
Hope it clears a bit the issues.
Best regards
Radek
________________________________________________________________________
Radoslaw Krzywania Network Research and Development
Poznan Supercomputing and
radek.krzywania@man.poznan.pl <mailto:radek.krzywania@man.poznan.pl> Networking Center
+48 61 850 25 26 http://www.man.poznan.pl
________________________________________________________________________
*From:*Jerry Sobieski [mailto:jerry@nordu.net] *Sent:* Tuesday, July 26, 2011 6:39 PM *To:* radek.krzywania@man.poznan.pl *Cc:* nsi-wg@ogf.org *Subject:* Re: [Nsi-wg] GEANT topology for demo
Hi Radak-
Looking at the tif file of the GEANT topology...
- Why do you have a 1G switch matrix defined with no ports? and why do you have a Node with two Switchmatrix objects? - Why have you defined hasPort relationships from both the Node object and the Switchmatrix object? This seems redundant...? Would the proper structure have the Node->Switchmatrx->Port(s) ??
Thanks! Jerry
On 7/26/11 9:46 AM, Radek Krzywania wrote:
Hi all, In the attachment there is a first draft of the GENAT topology. There will be small changes to that, i.e. to the URIs. Some of the ports are under implementation and the final numbering is missing (x/y instead). For some external ports I guess we need to agree on URIs: - we have connection to Internet2 - will that be also shown on Internet2 topology? If so, what would be the URI at Internet2 site? - What would be an URI at port going to GN from Netherlight for this demo?
Best regards Radek
________________________________________________________________________ Radoslaw Krzywania Network Research and Development Poznan Supercomputing and radek.krzywania@man.poznan.pl <mailto:radek.krzywania@man.poznan.pl> Networking Center +48 61 850 25 26http://www.man.poznan.pl ________________________________________________________________________
_______________________________________________ nsi-wg mailing list nsi-wg@ogf.org <mailto:nsi-wg@ogf.org> http://www.ogf.org/mailman/listinfo/nsi-wg