cancel
Showing results for 
Search instead for 
Did you mean: 

Routed topology resiliency

Routed topology resiliency

nameduser
New Contributor
In a routed topology the wireless controller acts as a router. The controller uses untagged VLAN IDs to assign routable addresses (Gateway and Mask).

As opposed to bridge at controller where the controller acts as a bridge and forwards to a switch where the gateway exists, in a routed topology, the gateway exists on the wireless controller itself and advertises the network using OSPF to a connected switch.

With multiple controllers spanning across sites using bridge at controller, as the gateway exists on a switch and not the controller itself, there must be a way to protect against failure to reach the gateway. This can be achieved by tagging a VLAN across links between where the switches connect to the controller and each other. In effect spanning the layer 2 domain across sites. By using VRRP, the failure of a switch or controller allows the gateway to still be reachable.

With a routed topology and multiple controllers spanning across sites, mobility must be used to ensure clients maintain their IP address.

If controller A has a routed topology with gateway 1.1.1.1 and controller B has a routed topology with gateway 2.2.2.1, when controller A fails, does controller B take over the gateway for 1.1.1.1? therefore the gateway is protected at layer 3 using mobility? In effect the controllers operate in HA fashion sharing configuration and controller B will then advertise the 1.1.1.0 network?

Also is the difference between standard mobility and central mobility the route the traffic takes? In standard the traffic will exit the local controller, in central it exits only the central controller?

What happens in regards to roaming between access points on different controllers in terms of which controller the traffic exits? I assume since both controllers are up then the network can't be advertised out both controllers, so client A connected to controller A advertising 1.1.1.0 goes out controller A, when it roams to an AP on controller B, since it's not advertising the 1.1.1.0 route, it will route the traffic back to controller A via its tunnel and traffic exits at controller A?

Are the routable topologies only routed via the management topology in terms of OSPF adjacency? an OSPF adjacency being formed between an IP on connected switch and the management topology IP address?
0 REPLIES 0
GTM-P2G8KFN