Network Bandwidth Calculator and Throughput Capacity Planner
Estimate how much internet or network bandwidth you need in Mbps or Gbps. Enter users and application workloads, including separate download and upload demand, then account for concurrency, protocol overhead, peak traffic, and target utilization to size an ISP, WAN, SD-WAN, or VPN link.
1. Describe your traffic
Start with a scenario
For each activity, estimate people, the percentage active at once, and bandwidth per active person.
Use this mode when monitoring or application data already gives you demand per user.
2. Planning assumptions
3. Capacity recommendation
—
Advertisement
Methodology and formulas
- Average workload demand:
users × concurrency ÷ 100 × Mbps per active user, summed separately for download and upload. - Busy-hour peak:
average demand × peak factor. - Traffic on the wire:
peak demand × (1 + overhead ÷ 100). - Required link capacity:
on-wire demand ÷ (target utilization ÷ 100). - Circuit recommendation: the smallest common tier at or above the larger independently sized direction.
Concurrency describes how many users perform an activity simultaneously; it avoids assuming every user is active at once. Peak factor covers busy-hour bursts above that concurrent average. Overhead represents headers, encryption, tunneling, retransmissions, and management traffic. Dividing by target utilization reserves headroom instead of planning to saturate the connection.
Advertised bandwidth is a nominal rate, while useful throughput may be lower because of protocol overhead, ISP contention, packet loss, device limits, and traffic shaping. This model is a first-pass capacity estimate. It does not model Wi-Fi airtime or channel contention, latency and jitter, packet loss, QoS queues, ISP oversubscription, application-specific burst behavior, or SD-WAN failover states. Validate important designs with measured busy-hour traffic and vendor requirements.
Bandwidth planning reference tables
Typical demand per active user
| Activity | Download | Upload | Notes |
|---|---|---|---|
| Web browsing | 0.5–1.5 Mbps | 0.1–0.3 Mbps | Average active use; page loads burst higher |
| SaaS / cloud apps | 1–3 Mbps | 0.3–1 Mbps | Varies with media and file transfer |
| HD video call | 2–4 Mbps | 2–4 Mbps | Allow low utilization for latency |
| VoIP call | 0.1 Mbps | 0.1 Mbps | Add codec and packet overhead |
| File sync / backup | 1–10+ Mbps | 2–25+ Mbps | Rate-limit or schedule bulk traffic |
Utilization and overhead guidance
| Network / traffic | Planning value | Reason |
|---|---|---|
| Voice/video or critical WAN | 50–65% utilization | More burst and latency headroom |
| General business internet | 65–75% utilization | Balanced cost and responsiveness |
| Predictable bulk transfer | 75–80% utilization | Higher queueing may be acceptable |
| Ethernet/IP overhead | 3–8% | Framing and protocol headers |
| VPN/tunnel allowance | 10–20% | Encryption, encapsulation, MTU effects |
Common link sizes
100 Mbps · 250 Mbps · 500 Mbps · 1 Gbps · 2.5 Gbps · 5 Gbps · 10 Gbps. Provider products vary; confirm whether quoted speeds are symmetric.
Network bandwidth calculator FAQs
How much bandwidth do I need for 50 users?
As a starting point, 50 SaaS-heavy office users at about 65% concurrency plus calls and cloud activity often justify 250–500 Mbps. Select a matching preset, then adjust the workload rows for a tailored answer.
How much bandwidth do I need for 100 users?
A typical 100-user office often lands between 500 Mbps and 1 Gbps, but video, backup, guest, and upload demand can shift it substantially. This calculator sizes both directions from the actual mix.
What Mbps per user should I assume?
Prefer activity-specific figures: light web use may average 0.5–1.5 Mbps download per active user, while an HD video call commonly needs 2–4 Mbps in each direction.
What is a safe utilization target?
70% is a useful general default. Use 50–65% for critical or latency-sensitive links and up to 75–80% for predictable bulk traffic.
Should upload and download be calculated separately?
Yes. Workloads and ISP plans are frequently asymmetric, so the tool independently calculates the required downlink and uplink.
How does VPN overhead affect bandwidth?
Encryption and tunneling add headers and can cause fragmentation or retransmission. A 10–20% allowance is common, but protocol and MTU determine the real cost.
Is bandwidth the same as throughput?
No. Bandwidth is nominal link capacity; throughput is useful traffic delivered after overhead, contention, loss, and other constraints.
How much headroom should a business internet connection have?
Keeping expected busy-hour load at 60–75% leaves 25–40% headroom for bursts, growth, and uncertainty.
Can this size a WAN or SD-WAN link?
Yes, as a first-pass estimate. Also test latency, loss, failover capacity, QoS, and provider contention before deployment.
Ownership, privacy, and limitations
Author:
Starlight Robotics Tools Team
Technical review:
Starlight Robotics editorial review
Last updated:
Privacy:
Inputs and calculations stay in your browser.
Methodology note: Transparent workload, concurrency, overhead, peak, and headroom formulas are shown above. Limitation: results are planning estimates, not a service guarantee; validate with monitoring, application SLAs, provider terms, and equipment specifications.