Redundancy and failover for telephony

Map the dependencies

Telephony depends on more than the PBX. Go through each dependency and ask what happens if it disappears.

Dependency	What to ask
Power	Uninterruptible power supply (UPS) and backup power for servers, the network, the switches that power the phones, SBCs, DECT base stations and the operator’s equipment on your premises. How long does it last, and are the batteries maintained?
Network	Redundant switches and firewalls? Separate routes between buildings and data centres?
Operators	How many SIP trunks (the operator’s phone lines and numbers over the internet), from which operators and over which physical routes? Can the operator reroute your numbers during an outage?
SBC	Redundant SBCs at different sites? Who manages them?
PBX servers	Clusters or redundant nodes in different data centres? What carries over at failover: queues, recording, reporting and integrations?
Licences	Does the system work if the licence server cannot be reached, and for how long? Check for each platform and version.
Cloud services	Teams, sign-in and suppliers’ cloud services: what works locally without them?
DNS and time	SIP often needs working DNS, including from the backup site. Time synchronisation is needed for certificates and logs.
Certificates	When do they expire, and who renews them? An expired certificate causes the same outage as a fault.
Integrations	Directory, CRM, patient record integrations, alarm systems and recording.
People and suppliers	On-call staff, access rights, contact lists and the supplier’s escalation path.

Typical backup routes for redundant telephony

• Two SIP trunks from different operators, ideally over different physical routes and different access types, such as fixed and mobile.
• Redundant SBCs and PBX nodes at different sites.
• Local survivability in branch offices and hospitals, so they can still make calls when the link to the central site is lost.
• Call forwarding by the operator to a backup number or to mobiles when the PBX does not answer.
• Phone subscriptions that work independently of the PBX, or another PBX as a backup.
• A change in how you communicate: calls and chat over the internet instead of telephony, or Wi-Fi calling instead of the mobile network.
• Manual backup procedures when nothing works, such as Rakel (the Swedish radio network for emergency and public safety services), messengers and home visits.

Mobile networks are a good backup but no guarantee. During major power cuts they come under heavy load, and capacity in wireless networks normally cannot be guaranteed.

How to test continuity

• Tabletop exercise: walk through a scenario, such as a power cut, an operator outage, a data centre outage or a cyberattack, with the business, IT and your supplier. Who does what, and who is informed?
• Planned failover test in a maintenance window: shut down a node, disconnect a SIP trunk, switch off an SBC or cut the link to a branch office. Measure the time to recovery and what stopped working.
• Test calls after every test: inbound to the PBX and queues, outbound, internal, callback, alarms to DECT and recording. The route for 112 is checked according to the plan agreed with the operator and SOS Alarm.
• Backup power: test running on battery and generator, not just that they start.
• After major changes: upgrades, new operators and server moves.
• Evaluate: shortcomings go into an action list and are followed up at the next test.

What to document

• Dependency map and network diagram: what depends on what.
• Numbering plan and routing, including in fallback mode.
• Continuity plan with backup procedures: when it is activated, who decides and what is done step by step.
• Contact list for operators, suppliers, on-call staff and your emergency response function, on paper as well.
• Contracts, such as your support contract for PBX and Teams: availability, response time, resolution time, maintenance windows and service reviews.
• Test reports and change log.
• Material for incident reporting within 24 and 72 hours, if you are covered by the Swedish Cybersecurity Act.

Also consider what contracted availability means. According to the Swedish Post and Telecom Authority (PTS), 99 per cent availability around the clock equals just over three and a half days of downtime a year.

What do the authorities say?

PTS guidance for organisations that use telecoms services and the business continuity guidance from the Swedish Civil Defence and Resilience Agency (MCF) say broadly the same thing: map your dependencies, set measurable requirements, have backup solutions and practise using them. No Swedish law says how often an organisation should test failover. For organisations covered by the Swedish Cybersecurity Act, continuity and crisis management are part of the requirements, but the Act does not prescribe any particular technology.

This guide describes general practice. How much redundancy you need is decided by your own risk analysis. Our offering for resilient telephony for mission-critical operations is described on a separate page.

Last updated 14 October 2026.