How to Design Two-Way Radio Coverage for a Data Center

Evertac Engineering Team · Updated November 2026

Data halls are among the hardest buildings to cover with radio: the outer wall, every row of racks and every containment door take a share of the signal, so handheld radios and mobile phones that work in the car park often fail inside. The reliable approach is a fiber-fed distributed system — one central signal source with remote units and low-power antennas inside the halls, power rooms and basements — planned for the whole campus at design stage rather than added building by building after handover.

Why radio coverage is hard inside a data center

Signal loss inside a data center adds up layer by layer. These are the typical ranges we plan for:

Every 10 dB of loss leaves one tenth of the signal power. After two or three of these layers, an outdoor base station or the public mobile network is no longer usable on the data hall floor. UPS rooms, switch rooms and generator rooms add electrical noise on top.

These are exactly the places where people need to talk when something goes wrong. Outage studies by the Uptime Institute have repeatedly found human factors and procedures that were not followed correctly behind a large share of serious outages. Radio coverage does not prevent faults by itself, but in the first minutes of an incident everyone on site needs to hear the same instruction and confirm it.

Design the campus, not the building

Large data center campuses are usually built in phases: one building in operation, the next under construction, a third still on paper. The common mistake is to buy a separate radio system for each building. Each one passes its own acceptance test, but the campus never becomes one system — calls drop between buildings, talk groups do not carry over, frequencies are applied for again and maintenance teams manage several platforms.

A campus design works the other way round:

1. Centralise the signal source. Place the base station and core in one building and plan channels, talk groups and management once for the whole campus.
2. Distribute over fiber. Use the campus fiber backbone to reach optical remote units (ORU) in each building. One remote unit typically carries 4–12 channels, so coverage capacity grows without a second core.
3. Cover each zone to its own conditions. Open halls are planned by area; dense rack zones and metal partitions need more, closely spaced antennas; long corridors, cable tunnels and basements are reached with distributed antennas fed from the same remote units.
4. Add buildings as nodes. When the next phase is built, it only needs new remote units and antennas — the core, numbering and talk groups stay the same.

Electromagnetic compatibility

IT equipment owners are right to ask whether radio coverage can interfere with servers and power systems. A distributed system helps here: instead of one high-power transmitter pushing signal through walls, many low-power antennas sit close to the users, so field strength near equipment stays low while coverage stays even. Antenna placement, output power and frequency should still be reviewed against the site's EMC requirements during design.

Spectrum in Southeast Asia: decide what really needs a private channel

Once data center programmes move across borders, spectrum becomes the first constraint. In Malaysia, frequency applications and approvals usually need several months of lead time; in Thailand, each additional frequency pair adds cost; in Singapore, the spectrum available for private radio is limited to begin with.

If every conversation runs on licensed private radio, spectrum dictates schedule, cost and future expansion. So the first design question is which communications genuinely need it:

The networks can be split, but the user's device should not be. Technicians should not have to choose a network in the middle of a fault. A multi-mode radio that runs DMR and 4G and selects the path automatically keeps one device and one way of working for every user.

After handover: run it as infrastructure

A radio system that is only discovered to be faulty when someone needs it is not data-center-grade infrastructure. Plan for day-to-day operation as well as coverage:

• Monitor the coverage system itself. Remote units and antennas report their status, and faults alarm to a specific device and location.
• See people before dispatching them. Indoor positioning puts responders on a map, so the first radio call is no longer "where are you?".
• Link patrols and work orders to the radio. Patrol routes, check-in points and voice records sit in one task record; an abnormal finding becomes a work order sent to the nearest technician with location and deadline.
• Keep operational data on site. Positioning, task and incident records can stay on the operator's own servers.

Common mistakes

• Treating radio as the last and cheapest purchase, after the building is finished.
• Buying one system per building instead of one architecture per campus.
• Testing only whether people can hear the control room, not whether the control room can hear them back.
• Putting every user on licensed private radio, then running out of spectrum at the next phase.
• No monitoring: the first sign of a failed remote unit is a missed call during an incident.

Design checklist

• Campus master plan: current and future buildings, fiber routes between them.
• Coverage zones: data halls, power and battery rooms, generator yards, loading bays, basements, stairwells, roofs.
• Which users and talk groups need private radio, and which can use 4G or Wi-Fi.
• Frequency plan and approval lead time in each country.
• EMC requirements for antenna placement and output power.
• Integration with security, BMS or DCIM and the incident process.
• Acceptance test method: coverage targets measured in both directions, per zone.
• Who monitors the system after handover, and how alarms are handled.

How Evertac approaches data center projects

Over the past three years Evertac has worked on data center programmes in Malaysia, Thailand, Singapore and Indonesia, from campus planning and coverage design to frequency applications, certification and expansion. Our DistriNET coverage systems distribute a central signal source over fiber to remote units and smart antennas, the PNR2100 radio works on DMR and 4G with automatic network selection, and the NetFlex platform handles device monitoring, positioning, patrols and work orders. See our data center solution for more.

FAQ

Can mobile phones replace two-way radio inside a data center?

Usually not on their own. Public mobile signal rarely survives the walls, racks and enclosures of a data hall, and operators of facilities hosting customer data generally want an operational voice channel they control. Mobile networks are useful for coordination between buildings and sites, alongside private radio.

Will the radio system interfere with IT equipment?

A distributed system uses many low-power antennas close to users rather than one powerful transmitter, which keeps field strength near equipment low. Antenna placement and output power should still be checked against the site's EMC requirements at design stage.

When should radio coverage be designed?

At the same stage as power and cooling. Fiber routes, antenna positions and equipment rooms are much cheaper to reserve in the design than to add after fit-out, and spectrum approval can take months in some countries.

How does the system grow when new buildings are added?

With a campus design the core stays in place; each new building adds optical remote units and antennas connected over the campus fiber, keeping the same channels and talk groups.

Do contractor teams need their own licensed frequencies?

Not necessarily. Temporary teams and contractor crews can use virtual channels over 4G or Wi-Fi, which does not consume private-radio spectrum or require an additional licence.