Skip to content
SD-WAN & MPLS

SD-WAN Is Only as Good as the Underlay: A Procurement Reality Check

Overlay controllers and policy engines get the demos — but latency, jitter, and failover behaviour still come from the circuits underneath. Here is how to buy underlay that matches SD-WAN promises.

SD-WAN vendors rightly emphasise application steering and zero-touch provisioning. What slides rarely show is a brownout on a cheap broadband underlay that the controller cannot fix because the path itself is degraded, not just mis-prioritised.

Enterprise procurement teams often split WAN underlay across three RFPs: fibre at HQ, broadband at branches, and LTE backup. Each award optimises locally; together they can produce asymmetric routing that breaks voice and video even when dashboards look green.

Carrier-neutral sourcing lets you normalise SLAs and diversity across sites without running separate negotiations in every city. Capcon aligns underlay choices with how your SD-WAN actually measures path quality — not just headline Mbps.

Before the next refresh, baseline three things per site: measured latency to your top five SaaS targets, failover time between primaries and backups, and whether backup paths share last-mile infrastructure with primary. If two of three are unknown, the SD-WAN rollout is flying blind.