Continuing from my previous and first post on EVPN on Catalyst - Layer 2 we’re looking in to multihoming with EVPN on Catalyst. This post will only show the multihoming related configuration. For full EVPN L2 config check out my first post. Redundancy is essential to ensuring business continuity in today’s infrastructures. EVPN has the ability to provide link and node redundancy for both devices and networks. It comes in two flavors:

For decades we’ve used MPLS layer 3 VPNs for building overlays in the enterprise networks. Although MPLS layer 3 VPN is a proven technology it lacks important security features that are requested of todays networks. Also, building layer 2 overlays using MPLS never scaled and worked best for P2P connections with no involvement of a control plane. EVPN addresses both these shortcomings of MPLS services and this post focusses on the workings of an EVPN layer 3 service using a simple example.

Things change. Especially in the realm of infrastructure. We tend to continuously focus on optimizing and redesigning our networks to support the business goals. Or you might simple need to repurpose hardware. Whatever the reason might be we need to know the “how” and “what” of such changes. Take StackWise Virtual, for example. If we want to break up the StackWise Virtual switch pair, you can find this official documentation:

Fabric-based networks are no longer exclusive to the data center or service provider networks. Today we use fabric-based networks all over to provide the network services needed to support our business. I like to define a fabric-based network by two things: A network of devices (typically switches) that optimally inter-connect in a strict uniform way (think physical topology here) A robust network built to support L2 and L3 services on top of it using the concepts of underlay and overlays Roughly speaking the underlay is nothing but a L3 network responsible for providing reachability between the devices that hold our endpoints (be they users or servers).