public Bridging domains

Hello, I see that there is a model for l2 network instances but I can't find information related to bridging domains (IRB, vlans, mac address table, etc.). Is that missing or did I overlooked something?

Thanks! David

Sep 17 '16 19:09 dbarrosop

hi David, much of this is in the interfaces model:

IRBs are routed-vlan interfaces (defined in the vlan model and augmented as a virtual interface type into interfaces
L2 VLANs are switched-vlans under Ethernet interfaces (global VLANs can be defined and referenced under .../network-instance/vlans/)
MAC address table (ARP entries) are under subinterfaces/subinterface/ipv[46]/neighbors

let us know if you have other questions or feedback. thanks.

Sep 17 '16 22:09 aashaikh

I am failing to find the equivalent of the "mac address table" construct. subinterfaces/subinterface/ipv[46]/neighbors seems to be the "ARP table" but I am not looking for that one. I am looking for the mapping between mac addresses and ports for a given broadcast domain (vlan).

Thanks for the pointers, much appreciated.

Sep 17 '16 22:09 dbarrosop

Right, sorry, what I mentioned is indeed the ARP table entries. I'm not sure we have the equivalent of the mac addr table exactly --- there is a mac-address property on each Ethernet interface, which in turn is tied to a VLAN -- so I think such a table could be built (or applied) from the model.

Sep 17 '16 23:09 aashaikh

But the mac address property is supposed to be the physical address of the interface, isn't it? Not a remote device connected to the same segment. To be able to work it should even be a list as you can have multiple mac addresses coming from the same port (a trunk, for example)

Some background, I am here on behalf on the project napalm-automation/napalm, not sure if you are familiar with it. We have been trying to abstract gathering data from devices using the existing tools we have today (textfsm, vendors APIs, netconf, etc). Until now we have been going our own defining with more or less success our own data models. However, we have decided we should investigate if it was feasible to start using OC/IETF models rather than our own so here I am : )

Thanks again for your help.

Sep 17 '16 23:09 dbarrosop

sorry, i misunderstood you were looking for configuration data related mac addrs and vlans. I think what you want is the FIB model? It reflects the mac addresses learned from remote systems -- see the match container in each table entry for the mac-address. We've just published that model recently and will be updating the network instance model to include it to enable multiple bridging domains / VSIs.

am definitely familiar with the work you guys are doing with napalm -- we're definitely interested in helping you use OC where possible (I believe @robshakir already working with you).

Sep 18 '16 00:09 aashaikh

I don't think that this is reflected in the FIB. The FIB will show the installed forwarding entries that there are, including characteristics such as the L2 encapsulation that might be used (if this is relevant). There isn't currently something that shows the learned MAC addresses within an L2 FIB. I'd expect that this would be something that we could add to the network-instance model (likely under FDB). It needs to be linked to the VLAN model too.

Sep 18 '16 00:09 robshakir

@robshakir, that sounds about right to me. Is there something I can do to help and get that into the model? I have little experience with YANG models but I will be happy to try to help.

@aashaikh, I am sure both projects can benefit of working closer. It's gonna take time until we get the vendors to implement OC on their boxes so providing a translation from native<->OC could be interesting as an interim step. We have some proof of concepts but nothing clearly defined so if you want to help out give your opinion it's the perfect time for that : )

Sep 18 '16 07:09 dbarrosop

Rob is right, bridging table is usually kept separately, and depending of whether it is qualified learning or not, it would be associated with a bridge or bridge/vlan

Regards, Jeff

On Sep 17, 2016, at 5:20 PM, Rob Shakir [email protected] wrote:

I don't think that this is reflected in the FIB. The FIB will show the installed routes that there are, including characteristics such as the L2 encapsulation that might be used (if this is relevant). There isn't currently something that shows the learned MAC addresses within an L2 FIB. I'd expect that this would be something that we could add to the network-instance model (likely under FDB). It needs to be linked to the VLAN model too.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.

Sep 18 '16 09:09 jefftant

So, is there something I can do to help moving this forward?

Sep 28 '16 07:09 dbarrosop

hi David, we went ahead and drafted a model for the mac table, and just completed our internal review -- should be merging / publishing this week. thanks!

Sep 28 '16 07:09 aashaikh

Oh, that's great!!! Thanks!

Sep 28 '16 07:09 dbarrosop

This issue is stale because it has been open 180 days with no activity. If you wish to keep this issue active, please remove the stale label or add a comment, otherwise will be closed in 14 days.

Jul 05 '24 02:07 github-actions[bot]