Comparing live pricing and terms from 400+ carriers and platforms
Multi-site SD-WAN migration is a cutover operating playbook: phased calendar, carrier coordination, parallel run, one owner after go-live. Not a weekend appliance swap.
Homepage line: a 12-site SD-WAN, 38 days end to end, one client, we do not publish who. Use it as proof that cutovers can be described—not as a guarantee for your estate. Lead times, builds, and notice windows decide the calendar.
Circuits, terms, notice dates, voice dependencies, and which sites cannot fail on a Friday. Demo wins do not replace a register. Tie to multi-site inventory and notice windows.
DIA, cable, fixed wireless, LTE—per address. Path diversity matters. Construction lead times dominate. On-net vs dig shows up here the same way it does on fiber quotes—same-building spread. Shop underlay across the market; do not assume the overlay vendor’s house circuit is best at every plant.
Prove apps, voice, and failover on a forgiving site before you bet the warehouse network.
Group by underlay readiness and contract dates—not by org chart convenience alone.
Keep MPLS (or legacy) until acceptance. Dual spend beats a site with no rollback.
Cut legacy after “done” is defined—and after notice/ETF math is clean.
| Role | Owns |
|---|---|
| Buckeye | Carrier PM, calendar, escalation, contract sequencing |
| Carrier / underlay | Circuit turn-up, last-mile, tickets |
| SD-WAN admin / MSP | Overlay policy, appliances, app validation |
| Site contact | Access, local testing, go/no-go signal |
First invoices lie. Tickets stall. Notice windows return. Stay is the point of an advisor cutover—not only the go-live week. See Stay and how we get paid.
Shop 400+ for transport and platforms. Contracts in your name. $0 advisory. Owner-led PM—Jonathan, not a rotating implementer. Coverage worldwide with a Midwest sweet spot for Ohio plants and warehouses. Related product page: SD-WAN & multi-site.
Overlay vendors will show a clean wave chart. Carriers will quote intervals that assume landlord access, power, and a lit building. Reality inserts special construction, delayed cross-connects, and a voice dependency nobody put on the slide. That is why inventory and contracts come before appliances, and why Buckeye treats underlay Compete as a first-class workstream.
Parallel run feels wasteful until the morning a wave fails acceptance. Temporary double spend is a controllable cost. An unplanned plant outage is not. Define “done” per site: apps, voice, failover test, and who signs acceptance. Only then release MPLS or legacy broadband—and only in line with notice and ETF math.
The 12-site / 38-day example proves a cutover can be described without a logo wall. It does not prove your estate fits that window. If fiber builds are in play, think in months. If sites are on-net and contracts align, compressed calendars happen. Start from the register either way.
After go-live, Stay is invoice truth and escalation continuity. SD-WAN projects that end at “turned up” leave finance reconciling mystery MRC and IT rediscovering account numbers at 2am. Keep one owner.
Manufacturing and distribution estates rarely fail because the SD-WAN GUI was confusing. They fail because Site 7’s underlay slipped, voice still rode the old path, and nobody owned the carrier bridge call. Buckeye’s Midwest sweet spot means we see a lot of those demarcs. Coverage is worldwide when your map is wider—method stays the same.
Align waves to maintenance windows that ops can actually staff. Do not schedule four plants on the same night because a project plan looked tidy. Keep site contacts named on the register. Keep rollback criteria written. Keep MPLS until acceptance is boring.
If life-safety analog still hangs off copper at a plant, do not bury that inside an SD-WAN project. Split it—link the copper life-safety page—and sequence it as its own compliance workstream. Overlay cutovers do not make fire panels into VoIP seats.
They fail when underlay is treated as a footnote. They fail when MPLS is cut on day one. They fail when voice dependencies are discovered during wave three. They fail when notice windows and ETFs were never on the project plan. They fail when the overlay vendor, the carrier, the MSP, and the advisor each think someone else owns the bridge call.
Buckeye’s cutover PM exists to prevent that ambiguity. We compete transport and platforms, sequence the calendar, keep contracts in your name, and remain after go-live. We do not replace your SD-WAN administrator or your MSP. We make sure carrier reality and contract reality show up in the wave plan before anyone ships appliances to a plant that is still waiting on a lateral.
Use the anonymized 12-site / 38-day story as proof that cutovers can be narrated without logos. Use inventory and carrier intervals to build your dates. If Jonathan confirms more boundary detail later, we will add it—until then, the TODO comments mark the proof gaps instead of inventing them.
Two circuits from one trench are not diversity. Cable plus LTE can be diversity. DIA plus a second provider on a separate lateral can be diversity. SD-WAN makes policy clever; it does not invent a second physical path. Compete the underlays with eyes open, then let the overlay do its job.
Voice and life-safety dependencies deserve explicit rows on the cutover checklist. A plant can pass an app test and still fail when the analog fire path or the SIP trunk was assumed. Split those workstreams. Do not hide them inside “network refresh.”
After acceptance, Stay includes invoice audit because SD-WAN projects spawn new MRC lines, credits, and partial months that AP will pay without reading. It includes escalation because hybrid periods create tickets that bounce between overlay and underlay vendors. It includes the renewal calendar because today’s underlay win becomes tomorrow’s notice window.
Book Jonathan with the site list, MPLS end dates, and any overlay shortlist. The first honest output is a calendar shaped by carrier intervals—not a go-live week chosen for slide aesthetics.
That list is a starting point, not Jonathan’s final anonymized checklist. Where proof is missing, HTML TODOs remain rather than invented metrics.
Carrier-direct SD-WAN bundles can look simple: one seller for overlay and underlay. Simplicity is not the same as best-on-net at every address. Independent Compete often yields a chosen overlay plus mixed underlays, with Buckeye owning the multi-carrier PM calendar. Ask any single-vendor pitch who escalates when their underlay miss slips your wave—and whether you can swap transport without rewriting the whole stack.
ETF and notice alignment is not optional project hygiene. A beautiful wave plan that ignores contract end dates becomes an expensive surprise. Build the cutover calendar from the register’s notice-due column, then layer technical readiness on top. That order prevents finance and network from running two different projects by accident.
$0 advisory. Residuals from providers. Contracts in your name. Owner-led PM. That is the Buckeye shape of an SD-WAN cutover—process first, feature tours second.
Related reading: multi-site inventory, fiber quote spread for underlay economics, 90-day notice calendar for contract gating, and the process page for Talk/Compete/Stay. Product overview remains on sdwan.html. This page is the cutover operating playbook those pages should link when the buyer moment is migration, not feature comparison.
Send the site list. Bring MPLS end dates and any overlay shortlist. Thirty minutes with Jonathan is enough to know whether your calendar is measured in weeks or months—and what inventory is still missing before anyone ships appliances.
Thirty minutes with Jonathan. Bring MPLS end dates and any SD-WAN shortlist. No deck.