Gap Density

What It Measures

Gap Density is the companion to Burst Density. While burst density measures the fraction of time spent in a degraded loss state, gap density measures the fraction of time spent in the gap state — periods where the network is delivering packets normally, below the burst threshold. A value of 0.90 means 90% of the call was clean delivery.

Together, burst density and gap density describe the loss topology of a call: a call with high gap density and low burst density experienced mostly clean delivery with short, infrequent loss events. A call with low gap density experienced the opposite — most of the call was in a degraded condition.

Why It Matters

Gap density exposes what happens between loss events. Two calls can share the same average packet loss rate but have completely different gap densities depending on how loss is distributed. Understanding gap density tells you whether your network is genuinely recovering between incidents or limping from one burst to the next.

For SIP infrastructure validation, gap density is a critical health signal:

• A network with high gap density (above 0.90) can absorb occasional burst events without catastrophic quality impact — codec concealment works within gaps.
• A network with low gap density (below 0.70) never gives the endpoint time to recover. The jitter buffer is perpetually stressed, and concealment becomes audibly degraded.

This matters especially during load testing: as concurrent call count increases, gap density often drops before average loss rises. It's an early warning signal that your infrastructure is entering a degraded operational regime before users start complaining.

Thresholds

How to Fix It

1. Read alongside Burst Density. Gap density and burst density should sum close to 1.0. If the total is significantly less, the RTCP-XR implementation has a threshold counting discrepancy — worth flagging to your SBC vendor.
2. Monitor trends under load. During stress tests, plot gap density as CPS increases. The point where gap density starts dropping is your network's effective capacity ceiling.
3. Reduce burst sources. Most gap density degradation originates from the same causes as burst loss: queue overflow, route instability, and aggressive firewalls. Fix burst density first; gap density will recover.
4. Evaluate Gap Duration. Low gap density combined with short gap duration means bursts are frequent and close together — a more severe condition than infrequent long bursts.

Related Metrics

• Burst Density — Companion metric: fraction of time in the burst loss state
• Gap Duration — Average duration of each gap period in milliseconds
• Burst Duration — Average duration of each burst event
• MOS Score — Gap density directly influences MOS through its effect on burst-aware E-model impairment