Model link capacity with overhead and control traffic. Allocate per class using weights, percentages, or demand. Get actionable QoS budgets for every network change.
Use this sample to understand typical class mixes on a 100 Mbps WAN link.
| Class | Traffic | Method | Value 1 | Value 2 | Value 3 | DSCP |
|---|---|---|---|---|---|---|
| Voice | RTP calls | voice | 64 kbps | 20 ms | 18 bytes | EF (46) |
| Video | Conferencing | percent | 25% | 40% peak | — | AF41 (34) |
| Business Data | Apps | weight | 6 | — | — | AF21 (18) |
| Best Effort | Web and updates | weight | 3 | — | — | BE (0) |
QoS policies fail most often when planners allocate from the physical link rate instead of the usable scheduler rate. This calculator first subtracts overhead and control traffic, then applies a safety margin so queues run with headroom. On a 100 Mbps link with 8% overhead, 2% control reserve, and 5% safety, the schedulable rate is about 85.5 Mbps.
Voice is sensitive to jitter, so it is usually treated as a reservation. The estimator uses packets-per-second derived from packetization time and adds RTP/UDP/IP plus layer‑2 bytes to payload. A 64 kbps stream at 20 ms packetization produces 50 packets per second, and headers can add meaningful overhead at small payload sizes.
Percent reservations provide predictable minimums for critical traffic such as conferencing or transactional systems. Weighted sharing is better for elastic classes, because unused capacity is automatically redistributed by weight. A common pattern is reserving voice and video, then splitting remaining bandwidth between business data and best effort.
Demand mode helps when teams have measured throughput targets. Enter requested Mbps for each class and the tool distributes only the remaining bandwidth proportionally. This approach highlights gaps: if total demand exceeds remaining schedulable capacity, each class receives a reduced share, revealing the need for a higher circuit rate.
Many designs pair minimum allocations with a maximum cap. In this calculator, Value 2 can act as an optional peak percent ceiling against schedulable bandwidth. For example, capping video at 40% prevents a large meeting from starving business applications while still allowing video to use spare capacity when available.
The Plotly bar chart visualizes allocated Mbps per class so reviewers can spot imbalances quickly. Use the CSV to archive change requests and the PDF to attach a readable summary to tickets. Always validate final settings with real traffic captures and device queue statistics after deployment.
Use a realistic estimate for tunneling, encryption, and framing. For many WAN links, 5–15% is common. Measure if possible using interface counters and capture sizes.
Overhead accounts for bits on the wire. Safety margin reserves extra headroom to reduce queue saturation, latency spikes, and jitter during bursts or microcongestion.
It is a planning estimate based on packetization and header bytes. Real usage depends on codec, silence suppression, and encapsulation. Validate with call traces and device QoS counters.
Use percent for hard minimum guarantees and compliance targets. Use weight for elastic traffic that can share leftover capacity dynamically without fixed reservations.
Demand mode distributes only remaining schedulable bandwidth using your requested Mbps values. It is useful for comparing measured needs to available capacity in a change review.
Rounding can push summed allocations over the schedulable value by small amounts. Switch rounding to “No rounding” if you need strict arithmetic alignment for configurations.
Important Note: All the Calculators listed in this site are for educational purpose only and we do not guarentee the accuracy of results. Please do consult with other sources as well.