Would anyone be able to give me a little advice on my options for assigning IP Pools to clients. This seems to be my sticking point with SDA.
We have a brownfield migration (collapsed cores, L2 to edge, VLANs spread across access hubs and we must retain current IP addressing). If I migrate an access hub to SDA, and it has a number of differrent DHCP clients in different VLANs o nit, then what options do we have for the correct IP pool to be assigned to the clients before DHCP happens (so they are assigned an IP from the correct scope)?
My understanding is, like in trad networks, the DHCP scope is chosen based on the source IP of the DHCP request at the DHCP server (the IP of the client's gateway SVI). So, can you have multiple SVI's (anycast GW) on a single SDA edge switch? and if so, how would the correct SVI's be assigned to the correct clients before DHCP happens?
So for example we have a switch with the below port configs and clients connected. Corresponding SVIs on the core. If I convert this switch to SDA access edge, how could I ensure that the vlasn10 subnet is allocated to the client connected to gi1/0/1, and the vlan20 subnet is allocated to gi1/0/2? Would that result in having two Anycast GW SVI's automagically configured onto the switch by DNA?
gi1/0/1 - vlan 10 - 192.168.10.0/24
gi1/0/2 - vlan 20 - 192.168.20.0/24
Thanks for your time, if I can get an answer to this, it would be an enormous help!
Thanks for the reply. I have read that article previously, and I think I get how traffic moves through an SDA fabric. I think I have my head around everything except this one thing I seem to be missing and can't find an answer for (possibly because it's very obvious!).
Before the clients are registered, when you physically connect a client to a fabric switch, what determines what IP pool that client is assigned to. And in what order does this happen. Is it something like this?:
1 Client is connected to a fabric edge switch 2 Client gets assigned to an IP pool 3 Switch assigns the SVI for that IP pool 4 DHCP is then handled (DHCP now see's the SVI IP as the source ip of the DHCP request and therefore can assign from correct scope)
My misunderstanding comes at the point DHCP happens. In my mind:
a The SVI has to be assigned for the client to get an IP from the correct DHCP scope b The fabric edge switch needs to know what IP pool the client belongs to before it can assign or use the correct SVI for DHCP relay.
Really hope someone can clarify the order of things above?