Designing for Voice over Wi-Fi: Getting It Right the First Time

https://www.linkedin.com/pulse/designing-voice-over-wi-fi-getting-right-first-time-de-oliveira-pqvpe 

Voice over Wi-Fi isn’t just another SSID, it’s one of the most demanding wireless use cases we face as engineers.

Over the years, I’ve had countless conversations with customers who assume it’s as simple as “just adding voice” to their existing Wi-Fi.

But those of us who’ve spent time in warehouses, hospitals, or manufacturing sites know that voice has its own rules.

Coverage Is No Longer Enough

Traditional WLAN designs focused on signal coverage, ensuring -67 dBm was available across the floor plan. But for VoWiFi, coverage is just the beginning.

We must also deliver continuity, consistency and controlled interference.

Key design targets I follow:

• Primary coverage: -67 dBm or better at the user device.

• Secondary coverage: ideally within 3 dB of primary (especially for roaming handsets).

• Co-channel interference: keep overlapping cells below -85 dBm. These aren’t arbitrary figures, they directly influence call stability, roaming performance and the ability to maintain sub-50 ms handoff times.

Design for Mobility, Not Portability

Data networks can tolerate session resets. Voice can’t.

When you’re designing for mobile VoIP clients, whether DECT-over-Wi-Fi, SIP handsets, or push-to-talk tablets, you’re designing for mobility, not portability.

Every roam must happen without dropping a single packet of audio.

Fast and secure roaming protocols like 802.11r/k/v should be validated with your handset type and Fast Secure Roaming should be tested both in open and authenticated SSIDs before production.

Latency, Jitter and Packet Loss

Voice is sensitive to three things:

• Latency: < 50 ms end-to-end.

• Jitter: < 5 ms variation between packets.

• Packet loss: ideally < 1 %. These metrics must be verified on both wired and wireless segments. The RF layer can’t compensate for congestion or mis-tagged QoS queues further upstream, so ensure end-to-end QoS is implemented, from handset to core.

Quality of Service and Airtime Efficiency

Wi-Fi doesn’t guarantee absolute priority; it offers statistical advantage.

For that reason, ensure WMM is enabled and traffic is correctly mapped to Voice AC (UP 6). Separate SSIDs for voice and data are advisable and especially when you can dedicate 5 GHz channels for voice traffic only.

In multi-radio environments, keep voice and data on separate frequencies to minimise collisions and contention.

Channel Planning and Spectrum Discipline

Channel planning is one of the hardest aspects to get right for voice deployments.

• Stick to 20 MHz channel widths in 5 GHz or 6 GHz bands as wide channels reduce available non-overlapping spectrum and increase CCI.

• Avoid DFS channels if your handsets have poor radar-event recovery.

• Keep transmit power moderate (around 10–13 dBm for typical enterprise deployments) to create balanced cell sizes and predictable roaming zones.

Roaming Validation

Test roaming under real-world conditions.

Validate handoff times under 50 ms while performing active calls. If the infrastructure supports Fast BSS Transition and key caching, confirm interoperability with each handset OS and firmware version.

Codecs and Call Quality

When possible, use high-quality, low-compression codecs such as G.711 or Opus. Compressed codecs may save bandwidth but often degrade quality and increase susceptibility to jitter on shared mediums.

Always verify that your call manager and handset firmware align on supported codecs.

Client and Infrastructure Readiness

Before rolling out VoWiFi, confirm your infrastructure can meet the demand:

• Validate PoE budgets and switch QoS queues.

• Ensure DHCP lease times and DNS response are fast as these delays compound during authentication.

• On the client side, test with your actual handsets and make a note of the make, model, OS and firmware version, as performance can vary dramatically.

Design Beyond the Office Floor

Voice doesn’t stop at desks.

Elevators, stairwells, warehouses, parking structures, all need coverage if users will move through them.

These are often the forgotten areas where calls drop, leading to frustration and lost trust in the wireless network.

Verification and Continuous Optimization

After deployment, validate under load.

Use tools like iPerf to test throughput and measure upstream/downstream jitter, RSSI and retry rates.

Document your results and re-test whenever firmware or channel changes occur.

Voice performance drifts quickly when environmental or configuration changes are made.

Final Thoughts

Voice over Wi-Fi isn’t plug-and-play, it’s a discipline that demands deliberate design, validation and maintenance.

A network that passes a coverage test doesn’t automatically pass a voice test.

When you design for voice, you’re designing for experience: clear, uninterrupted communication that lets users forget they’re even on Wi-Fi.

That’s when you know you’ve done it right.