05-08-2026 04:36 AM
Hi Experts,
I have started working on SD Access recently. Just curious about the failover mechanism here.
I am sharing the design scenario which is standard for most of the designs. We have SD WAN for WAN connectivity and SD WAN is managed by another vendor of the client. We have eBGP configured between Fusion nodes and SD WAN routers as shown in the following diagram.
Could you please help me to understand how the failover happens at SD LAN end when the primary WAN link fails and WAN traffic moved on the secondary link? How exit traffic from LAN end to WAN routers will failover?
05-08-2026 04:54 AM - edited 05-13-2026 01:14 AM
hi
for simplicity let's assume your BNs are anywhere & BN#1 has better priority
then they will signal via LISP their availability as PETRs & ENs will use BN#1 as VXLAN termination point.
so while BN#1 has default route in arbitrary VN's RIB it's used as exit point. this default is propagated to BNs by FNs which BGP configuration is different story but essentially they will track routes from SD-WAN peers to make corresponding decision. in simplest case of primary link or SDWAN-peer failure FN#1 stops advertising default to BN#1. ENs start sending northward traffic to BN#2 (assuming it still has default route). That's it.
05-12-2026 11:55 PM
@Andrii Oliinyk Thank you for your time and the response !!
05-13-2026 01:15 AM
you are welcome)
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide