Server Bandwidth Calculator
Concurrent viewers times bitrate — the busy minute priced against the ports that must carry it.
Server Bandwidth Calculator
Results recalculate instantly on every keystroke. Nothing you type is transmitted.
What this result does not account for
- Linear sum at one minute — no burst smoothing
- Port constant 0.94 is TCP-over-Ethernet convention
In short: Five hundred concurrent streams at 3 Mbps each ask 1,500.000000 Mbps at the busy minute — two gigabit ports deep, because a port that says 1,000 delivers about 940 once TCP takes its share. Pinned at that minute for a day, the same load moves 16,200.000000 GB. Size for the busy minute, not the average: the average never made anyone add a port at 8 p.m.
Formula
Mbps = streams × bitrate ··· ports = Mbps ÷ (port × 0.94), rounded up
Concurrent load adds linearly: every viewer brings their bitrate to the same door at the same minute. The port constant is 0.94 — a gigabit port moves about 940 Mbps of real payload once TCP and IP headers take their cut, which is why port counting at the sheet rate under-buys by exactly that margin. The pinned-day figure is a ceiling thought, not a forecast: it answers what the load would move if the busy minute never ended.
Worked Example
- Enter viewers at the same moment.
- Enter each one's bitrate.
- Pick the port you would buy.
- Read the load, the port count and the ceiling.
Defaults: 500 streams × 3 Mbps → 1,500.000000 Mbps, 2 gigabit ports. Drive 100 streams at 8 Mbps → 800.000000 Mbps on one port with 140.000000 to spare.
Strengths & Limits Of This Model
Where this engine is strong
- Busy-minute framing, not averages
- Port count on the honest 940 figure
Where it stops
- Disk, CPU and origin uplink not modeled
Practical Use Cases
Stream sizing
ports needed for the premiere night
Launch capacity
the demo-day traffic question
Port upgrades
gigabit vs ten-gig, in viewers
Methodology & Editorial Standards
Computation runs in IEEE-754 double precision at full internal precision; rounding to two decimal places occurs strictly at the display layer, so no cumulative drift enters the result. All monetary outputs use accounting presentation — grouped thousands, two decimals, negatives in parentheses — so figures can be transcribed directly into a model or working paper. Division-by-zero and out-of-domain inputs return an em-dash rather than a misleading number.
This engine was reconciled against an independent reference implementation and hand-verified for the worked example above before release. Our full five-stage review process is published on the About Us page.
Disclaimer. This calculator is provided for informational and modelling purposes only and does not constitute financial, tax, legal, medical, or engineering advice. Verify all figures with a qualified professional before acting on them.
Server Bandwidth Calculator — 8 Expert FAQs
8 analyst-written answers to the questions practitioners actually ask — optimised for voice and answer-engine retrieval.
How do I size a server for concurrent streams?
Multiply concurrent viewers by each stream's bitrate and compare against what your ports really deliver. Five hundred viewers at 3 Mbps arrive together at 1,500.000000 Mbps, which is two gigabit ports once the 940-real figure is used. The arithmetic is linear; the judgment is picking the honest concurrency.
Why does a gigabit port only count as 940 Mbps?
Ethernet's line rate carries every frame's TCP and IP headers, preamble and gaps along with the payload — about 94% of a gigabit survives as real data in practice. Sizing ports at the full sheet figure plans a ceiling you cannot reach, which is the same promise-versus-delivery lesson the file transfer page teaches at 35 of 60 promised MB/s.
What bitrate should I plan per viewer?
Plan by what you serve: low-quality video sits near 1 Mbps, a solid high-definition stream 3–5, near-Full-HD approaches 8. Audio-only drops to fractions of a megabit. The streaming data page prices the same ladder in gigabytes per hour — this page prices it in simultaneous viewers.
What is the difference between this and a speed test?
A speed test measures one lane's moment; this page sums every lane that arrives at once. Your server's port is shared by all concurrent viewers, so the load that matters is the sum at the busiest minute, not what one connection can achieve alone. Both numbers matter; they answer different questions.
How many viewers can a gigabit server handle?
At 3 Mbps each, about 313 viewers before the 940 real megabit ceiling — the calculator runs the division for any bitrate you serve. Remember the server's own disk, CPU and the origin's uplink also vote; the port is usually the first wall, not the only one.
Should I size for average or peak traffic?
Peak — the busy minute is the only minute that can overflow a port, and overflow reads as buffering to the people watching. A service that peaks at eight evening hours needs the peak number on the port math, with the average reserved for billing conversations. The pinned-day read shows the ceiling's absurd scale on purpose.
Does a CDN change this calculation?
It splits it: the CDN's edge serves the cached share and your origin port only sees the miss share, so the same audience needs a fraction of the ports. The CDN bandwidth page prices that split in dollars. Run this page for the origin's remaining load after the hit ratio does its work.
Why does the pinned-day figure matter?
It is the honest ceiling of the busy-minute load: 1,500.000000 Mbps held for 24 hours moves 16,200.000000 GB, a number no real day reaches but one that bounds every billing surprise. Some hosts bill on transfer, so a load that fits the port can still blow the invoice — both walls deserve a number before launch.