Switching wholesale VoIP carriers sounds like a simple vendor swap — until the cutover window arrives and live traffic is on the line. Get it wrong and you get dropped calls, a week of overlapping invoices, or a support queue full of angry resellers. Get it right and the switch is invisible to every subscriber on your network.
Carrier migrations happen for good reasons: a route quality problem that never got fixed, a rate increase you didn't agree to, a compliance gap, or simply the arrival of a better-engineered partner. Whatever the trigger, the operational challenge is the same — you must move live SIP trunks, DIDs, and call volume from one termination network to another without a gap where calls fail to connect or bill twice.
Most failed migrations share the same root cause: teams treat the switch as a single event — flip the routing table, cancel the old contract, done — instead of a phased process. A hard cutover means every call in flight at the moment of the switch, and every call placed before DNS or gateway changes propagate, is exposed to the new carrier's untested routes with zero fallback.
The failure modes are predictable: destinations that tested fine in a sandbox but fail at production volume once the new carrier's routing tables get real traffic; SIP trunk re-pointing that breaks authentication or codec negotiation mid-switch; and a billing overlap where both carriers invoice for the same traffic window because nobody set a hard decommission date on the old route. Each of these is avoidable with a planned, staged migration rather than a single flip of the switch.
Map traffic by destination and volume so you know exactly which routes carry the highest risk if something goes wrong — a low-volume niche destination can tolerate a rougher cutover than your top revenue-generating corridor. Document current ASR, PDD, and MOS baselines for every major destination on the outgoing carrier so you have an objective standard the new carrier must match or beat before you shift meaningful volume.
A migration that touches meaningful call volume should never be compressed into a single weekend. Plan for a testing phase of one to two weeks, a parallel-run phase of two to four weeks, and a decommissioning phase that only begins once the new carrier has held quality benchmarks through at least one full billing cycle.
The single most important technique in a zero-downtime migration is running both carriers live at the same time, with the old route still handling full production traffic while the new one comes online alongside it. Configure your softswitch or session border controller with both carriers as active routes, using least-cost routing priority or percentage-based traffic splitting to control exactly how much volume flows to each side.
Start the new carrier at a deliberately small traffic percentage — 5 to 10% is typical — routed to a handful of lower-risk destinations first. This limits blast radius if something in the new carrier's routing or authentication setup is misconfigured, while still generating real production CDRs you can use to validate quality. Keep the old carrier's trunk fully provisioned and untouched throughout this phase; it is your instant fallback, not a decommissioned relic.
Never trust a new carrier's rate sheet or sales-provided quality claims — validate everything against your own traffic. Run structured test calls across every destination you plan to migrate, checking answer behavior, audio quality, DTMF passthrough, and fax or SMS compatibility if those services ride the same trunk. Compare the results directly against the ASR, PDD, and MOS baselines you documented during planning.
Set explicit go/no-go thresholds in advance — for example, ASR must stay within 5 percentage points of your baseline and PDD must stay under 6 seconds for three consecutive days — so the decision to expand traffic share isn't made on gut feel under deadline pressure.
Once the new carrier has proven itself on a limited slice of traffic, increase its share incrementally rather than jumping straight to 100% — a common cadence is 10%, then 25%, then 50%, then 100%, with a 24- to 48-hour observation window at each step. This gives you multiple checkpoints to catch a degradation before it affects your whole customer base, and each step is trivially reversible by adjusting routing weight back toward the old carrier.