Hybrid PSTN and SIP Deployments: Surviving the Overlap

Hybrid PSTN and SIP Deployments: Surviving the Overlap

Almost no organisation transitions from PSTN to SIP cleanly. Numbers arrive in stages, contract terms extend, a branch that cannot get SIP refuses to move, and a carrier circuit stays in place because the devices on it have not been replaced. What was meant to be a two-year transition becomes a steady state, and the estate becomes more complicated than either a pure PSTN or a pure SIP design ever was. The complication is not inherently harmful — a mixed estate offers a form of natural resilience — but the benefits only appear if routing between networks was designed deliberately. Left alone, hybrid estates accumulate ambiguous routing rules that produce calls to the wrong network, wasted capacity, and an upgrade path nobody can describe.

Why the Overlap Lasts Longer Than Planned

The reasons a PSTN estate persists are mostly commercial and operational rather than technical. Contract terms may run for years. A site may have equipment whose replacement is not justified by the volume of traffic. Regulatory or customer requirements may in some cases dictate how certain calls are routed. A carrier may offer a SIP service with a clause that makes it unattractive. Recognising this early prevents the design error of assuming the PSTN side will disappear, and it prevents capacity planning that only counts the SIP side.

A second reason is organisational. Adding SIP usually lands on the IT side while PSTN lines remain under facilities or a legacy contract owner. When two groups own halves of the same voice estate, each optimises its own part, and the routing between the two halves gets defined by habit rather than by design.

Designing Routing Between the Networks

The essential design question in a hybrid estate is which network carries which call, and how the system decides. There are several defensible approaches. Trunk-first routing sends calls to the SIP service and only uses PSTN when SIP is unavailable, which maximises SIP usage and keeps PSTN capacity as a reserve. Least-cost routing selects the cheaper or more appropriate path per destination, which is common where the two networks cover different destinations. Local-first routing keeps internal and nearby traffic on the legacy system while external calls move to SIP. None of these is universally correct, but choosing one and documenting it prevents the accumulation of ad hoc exceptions.

What must also be designed is the fallback behaviour. If SIP fails, the system should fail to PSTN predictably and quickly, and it should do so without requiring manual intervention. Conversely, if PSTN fails, the reverse should hold. Predictability is what matters: an estate where fallback behaviour depends on which device notices the fault first produces inconsistent call results that are very difficult to diagnose and very visible to users.

Number Plan and Trunk Management in a Mixed Estate

Number management is where hybrid estates most often go wrong. A number can exist on both networks, on the legacy system for a group that has moved to SIP, or on SIP while the legacy system still routes to it. Each of these states has consequences: duplicate number assignment produces unpredictable routing, and a number left on the legacy system consumes a PSTN resource that could be released.

• Maintain one authoritative number plan that shows where each number currently terminates
• Flag any number present on both networks, since assignment rules do not always prevent this
• Track numbers that could be released from PSTN as branches complete their migration
• Reconcile the number plan after every cutover, not only at project milestones
• Keep the mapping between extension, department, and network path documented
• Review the plan before any tariff or contract change, since routing may need to change with it

Operational Discipline During the Overlap

Running two networks indefinitely means running two monitoring regimes, two sets of troubleshooting knowledge, and two failure modes. Teams that accept this cost explicitly can manage it; teams that assume the overlap will be brief tend to under-invest in documentation and then discover the overlap is still in place when the original team has moved on.

A workable approach is to define what has to be true for each branch to be declared migrated, rather than defining what a migration looks like as a set of steps. When the criteria are outcome-based — calling works, numbering is released, documentation is updated, monitoring covers the site — branches finish on a consistent basis and the overlap shrinks for a visible reason.

Redstone Approach

Redstone can support hybrid estates where PSTN and SIP must coexist, including how RGW series gateways should be applied to keep analog devices and PSTN access available alongside SIP trunks on a Redstone IP-PBX. Redstone can advise on routing structure between the two networks, on how calling groups should behave when both paths are available, and on how survivability should be configured so that a fault in one network produces a predictable outcome.

For organisations still reducing a PSTN estate, Redstone can help plan the sequence — which sites move first, what stays on legacy circuits, and how to avoid a situation where the hybrid design becomes permanent by accident. Recommendations are advisory and based on the estate as it exists, not on a fixed migration template.