Communications / PBX planning

VoIP bandwidth planner

Size the upload and download capacity for concurrent G.711 audio calls using real packet/header arithmetic, not just the 64 kbps codec rate. Calculations stay in your browser.

Plan guest Wi-Fi voucher policy (offline worksheet)

Planning estimate, not a live network test. Assumes continuous bidirectional G.711 RTP audio, fixed packetization, no compression or silence suppression. It excludes SIP signaling, RTCP, SRTP, VPN/tunnel overhead, other traffic, jitter, latency and loss. Confirm your PBX/carrier settings and actual WAN performance before deployment.

Call and link assumptions

Use concurrent calls, not the total number of registered handsets.

RTP packetization

Smaller packets reduce packetization delay but increase header overhead.

Network layer

Overhead accounting

Includes header + FCS. Ethernet frame includes 18 bytes; physical wire includes 38 bytes before optional VLAN.

802.1Q VLAN tagging

Reserved capacity is withheld from the WAN before estimating call capacity; 25% reserve leaves 75% usable for this voice model.

Estimated media demand

Each direction is sized separately; do not double the call count on a full-duplex link.

Within modeled capacity

Per call, each direction

87.20 kbps

Upload media

0.872 Mbps

Download media

0.872 Mbps

Available voice capacity after reserve

Upload: 7.500 Mbps · 86 calls
Download: 18.750 Mbps · 215 calls
Upload margin: 6.628 Mbps
Download margin: 17.878 Mbps

Packet math

G.711 at 64 kbps; no assumed carrier bandwidth guarantee.

Voice payload: 160 bytes/packet
IP header: 20 bytes
UDP + RTP headers: 20 bytes
Link-layer and VLAN overhead: 18 bytes
Total accounted packet: 218 bytes
Packet rate, each direction: 50.000 packets/sec

Optical infrastructure planning

Plan fiber runs with editable attenuation, connector/splice loss and optional optical transceiver power budgets. Offline and assumption-based.

Open Fiber Optic Loss & SFP Budget Calculator

PBX planning / Phase 5B

Inbound call-flow scenario

Explore how assumed trunk capacity, ring-group capacity, queue limits and after-hours treatment affect incoming calls. All data stays in your browser.

Build a routing scenario

Guided setup or direct expert controls; no PBX connection.

Assumption · offline

Start with simultaneous inbound calls and the contracted external channel limit.

Step 1 of 3

Model provenance: vendor-neutral assumptions; RFC 3261 SIP is a signaling reference, not a guarantee of trunk, queue or voicemail implementation. Confirm licensed channels, IVR and queue policies with the selected PBX and carrier.

PBX planning / Phase 5B

Extension & SIP trunk capacity

Estimate simultaneous internal/external calls and SIP trunk channel needs from editable assumptions. This runs entirely in your browser; no PBX or provider is contacted.

Scenario assumptions

Illustrative examples only; replace with your own estimates.

Both percentages round up to whole calls. SIP trunk channels represent concurrent external sessions. DIDs are numbers and do not add channels. No vendor or carrier compatibility is asserted.

Estimated call demand

Planning scenario, not live busy-hour measurements.

Channels sufficient

Peak calls: 10
External calls: 7
Internal calls: 3
Trunk channels needed: 7
Channels provisioned: 8
Channel shortfall: 0

Modeled external-channel headroom: 1. Provider capacity is not verified.

3 DID numbers entered; this count does not increase simultaneous capacity.

Method: peak calls = ceil(extensions × concurrency %); external calls = ceil(peak calls × external share %). External calls consume one modeled trunk channel each; internal calls do not. Excludes carrier policy, PBX licensing, codec/transcoding, failover and call-flow routing. No live data or provider price is used.

Voice QoS / offline capacity planning

Voice priority and background traffic

Reuse the G.711 assumptions above to compare a proposed priority queue against competing WAN traffic. This does not configure QoS or predict call quality.

Editable planning assumptions

Example values only. Keep the original G.711 inputs above in sync automatically.

Compare a before-and-after scenario

Save a valid baseline, change any voice or QoS assumption, then inspect the difference. This is temporary in-tab planning, not a live network measurement.

RFC 4594 recommends EF for telephony. DSCP 46 is guidance, not a setting applied here. Priority requires trusted classification, queue policy and end-to-end validation; provider markings may be changed.

Modeled upload and download

Voice is full-duplex: compare each direction once, not twice.

Fits modeled budget

upload

Voice demand: 0.872 Mbps
Priority budget: 1.5 Mbps
Total with background: 5.872 Mbps
Usable link: 7.5 Mbps

Priority budget covers modeled voice. Aggregate link demand fits.

download

Voice demand: 0.872 Mbps
Priority budget: 3.75 Mbps
Total with background: 10.872 Mbps
Usable link: 18.75 Mbps

Priority budget covers modeled voice. Aggregate link demand fits.

Modeled simultaneous-call ceiling: 17 (queue and aggregate limits in both directions).

Derived arithmetic + editable assumptions, not measured performance. Excludes signaling, encryption/tunnel overhead, jitter, loss and router queue implementation.

References: IETF RFC 4594 telephony QoS, Cisco voice WAN design, and Microsoft Teams QoS guidance. Vendor-specific policies are not assumed. No PDF is offered; reports will use pdfcn after renderer verification.

Multi-WAN / Phase 5C

Failover capacity scenario

Test whether a backup WAN could carry your essential upload and download demand, then estimate the interruption from an assumed failover timer. No controller, ISP or Starlink account is contacted.

Hypothetical, not live. Example speeds are invented inputs for learning, not measured or advertised provider performance. This is active/passive only: links are never aggregated. Switching interruption excludes outage duration, recovery, DNS, NAT/session resets and application reconnection.

Link and demand assumptions

Choose an illustrative starter, then replace every number with your own planning data.

No historical failover events or uptime percentages are inferred. Enter only non-negative finite rates; all values stay in this browser session.

Scenario results

Derived capacity, not verified availability.

Backup meets modeled demand

Primary usable download: 150 Mbps
Primary usable upload: 37.5 Mbps
Backup usable download: 75 Mbps
Backup usable upload: 11.25 Mbps
Backup download shortfall: 0 Mbps
Backup upload shortfall: 0 Mbps

Primary: Meets entered demand

Backup: Meets entered demand

Assumed switch gap: 25 seconds/event

Projected switching gaps: 0.83 minutes/month; 10 minutes/year at the entered event count.

Method: usable capacity = link capacity × (1 − reserved %). Both directions must meet demand. Assumed switching interruption = detection + convergence; projected minutes = interruption seconds × hypothetical event count ÷ 60. This does not predict outage frequency, full annual downtime, availability, link quality, Starlink performance or billing. Starlink official plan documentation describes plan conditions, not this site's speeds. No PDF is offered; future PDF exports must use official pdfcn with verified bytes, pagination and redaction.

Sources and limitations

The reference G.711 20 ms RTP packet contains 160 bytes of voice payload and 40 bytes of IPv4/UDP/RTP headers, or 80 kbps before Layer 2 overhead. Ethernet frame accounting adds 18 bytes (87.2 kbps), while physical-wire accounting also includes the 8-byte preamble/SFD and 12-byte inter-frame gap (95.2 kbps). An 802.1Q tag adds 4 bytes per packet. These are accounting choices, not interchangeable measurements.

Cisco Unified Communications network infrastructure reference. No PDF export is implemented; future PDF reports must use approved pdfcn components after verified rendering and redaction.