Evidence taxonomy
Four number types, four different questions
- Vendor service requirement
- What a publisher lists for a named resolution, mode, or application. It applies to that scenario and can change.
- FCC household benchmark
- The FCC raised its fixed-broadband benchmark to 100 Mbps down and 20 Mbps up. It is policy context, not a per-app minimum or performance guarantee. Read the FCC report.
- Editorial planning target
- A transparent, versioned input for combining simultaneous household activities. It is conservative planning, not a universal pass/fail boundary.
- Measured diagnostic
- Evidence about one path, endpoint, device, load state, and time. It helps locate a bottleneck but does not guarantee every app.
Single source of planning data
Requirements dataset 2026.07.24
Updated 2026-07-24. Every value below is rendered from the versioned in-repo dataset. Latency columns are editorial planning ceilings, not universal strong, usable, or bad bands.
| Activity | Download | Upload | Ping ceiling | Jitter ceiling | Evidence label |
|---|---|---|---|---|---|
| Web browsing and email | 5 Mbps | 1 Mbps | 150 ms | 40 ms | Editorial planning baseline for an active browser and email session. |
| HD video stream | 8 Mbps | 1 Mbps | 100 ms | 30 ms | Editorial planning baseline informed by official service recommendations. |
| 4K video stream | 25 Mbps | 1 Mbps | 100 ms | 30 ms | Conservative editorial planning baseline informed by official service recommendations. |
| HD video call | 4 Mbps | 4 Mbps | 100 ms | 30 ms | Rounded planning baseline informed by official 1080p video-call requirements. |
| Remote work session (excluding video calls) | 10 Mbps | 3 Mbps | 100 ms | 30 ms | Editorial composite for concurrent work applications, not a vendor minimum. |
| Online gaming | 10 Mbps | 3 Mbps | 50 ms | 15 ms | Editorial gameplay baseline; title, platform, and route conditions vary. |
| 1080p livestream upload | 10 Mbps | 10 Mbps | 80 ms | 20 ms | Conservative editorial baseline informed by official broadcast bitrate guidance. |
| Cloud backup or large upload | 10 Mbps | 25 Mbps | 150 ms | 40 ms | Editorial target chosen for practical completion time rather than a service minimum. |
Household planning
Add simultaneous activity, then apply documented headroom
- 1. Count sessions that can overlap at the busiest realistic hour.
- 2. Add the per-session download and upload planning values.
- 3. Use the strictest latency and jitter ceiling among selected interactive activities.
- 4. The calculator adds 25 percent headroom after combining active sessions; this is a transparent editorial choice, not a vendor minimum.
- 5. Verify the current broadband label, plan policy, and real applications before buying.
Measured diagnostic
Decide which layer needs attention
- Capacity below the combined plan: identify which simultaneous activity fails and compare wired with Wi-Fi.
- Capacity fits, loaded latency rises: investigate queueing and competing transfers before buying more peak speed.
- Browser test fits, one app fails: inspect its endpoint, route, device, service status, and first-party diagnostics.
- Wide repeat spread: keep the variability; do not reduce it to one good or bad label.
First-party sources and versioned distribution
Source-checked July 24, 2026. The dataset exposes the evidence basis and source IDs for every activity row.
- https://swiftspeedtest.com/data/internet-speed-requirements/2026.07.24.json
- https://docs.fcc.gov/public/attachments/FCC-24-27A1.pdf
- https://help.netflix.com/en/node/306
- https://support.zoom.com/hc/en/article?id=zm_kb&sysparm_article=KB0058323
- https://help.twitch.tv/s/article/broadcasting-guidelines?language=en_US