Hello community,
We are migrating a campus from an existing MPLS campus with PE routers at each distribution block in the network. The original migration strategy was to acquire new hardware as intermediate nodes to be installed in the same location as the existing MPLS distribution switches. The plan was to move building by building from the old MPLS distribution switch to the new SD-Access intermediate switch. This is very expensive (running up two parallel networks with identical hardware side-by-side).
The existing switches are high-end and compatible as SD-Access underlay switches. The idea has been floated to reutilise the existing MPLS switches as SD-Access intermediate nodes, thereby reducing the amount of hardware required while running the old network transport and new SD-Access underlay.
The underlay routing protocol will not change (currently OSPF, will be OSPF in SD-Access underlay). As far as I'm aware, if we can get GRT layer 3 connectivity around the underlay, a GRT default route through the border and adjust the MTU for jumbo frames, we should have what we need for SD-Access to function in parallel to MPLS. I've got a simple image of the proposed migration below:
If you have any input or warnings, please advise!
Josh
