Network Test
Free Bufferbloat Test — A–F Grade 2026
3 phases · Idle + load comparison · Grades A–F · No plugins required
Bufferbloat makes fast internet feel slow during downloads. This 3-phase test measures your latency at idle, under download load, and under upload load, then grades your router A–F.
Baseline
Download
Upload
Results
We start by measuring your latency while the connection is completely idle. This is the best-case response time — we compare it to loaded latency in the next phases to reveal bufferbloat.
10 idle probes to establish your baseline ping before any load is applied.
We saturate your connection with a 25 MB download stream while probing latency every second. If your router has bufferbloat, you’ll see latency spike dramatically even though the download is working fine.
Downloading while probing latency every second for 8 seconds…
Upload bufferbloat is often the worst offender — cable and DSL connections have narrow upload channels that fill up fast. We POST 8 MB of data while probing latency to test your upstream buffer.
Uploading while probing latency every second for 8 seconds…
Why Fast Internet Feels Slow
Bufferbloat happens when your router has oversized data buffers. When the buffer fills up during a heavy download, all new packets — including gaming data, video call audio, and web page requests — sit waiting in queue. Your download speed looks fine, but everything else grinds to a halt.
A vs F — What the Grades Mean
| Grade | Latency Increase | What It Means |
|---|---|---|
| A | 0–5 ms | No bufferbloat. Ideal. Gaming and calls work perfectly even during downloads. |
| B | 5–20 ms | Minimal. Rarely noticeable. Good quality connection. |
| C | 20–50 ms | Moderate. Perceptible lag on video calls during file transfers. |
| D | 50–150 ms | Significant. Gaming becomes frustrating; choppy audio on calls during transfers. |
| F | >150 ms | Severe. Major cause of poor gaming and call quality. Action required. |
How to Fix Bufferbloat
Enable SQM (Smart Queue Management) on your router. Available in OpenWrt, pfSense, DD-WRT, and ASUS Adaptive QoS. Fiber connections (AT&T, Google Fiber, Verizon Fios) almost always grade A or B because symmetric speeds eliminate the upload bottleneck that causes bufferbloat on cable connections.
How This 3-Phase Test Works — and Why It’s More Thorough
Most speed tests just measure throughput. Bufferbloat is invisible to throughput-only tests — your download speed can show 400 Mbps while gaming is completely broken due to buffer saturation. This tool is specifically designed to reveal the gap between idle latency (what your connection can do) and loaded latency (what it actually delivers when in use).
Phase 1 — Why start with baseline?
Without a baseline, a loaded latency of 200ms is meaningless. If your idle ping is 8ms and load takes it to 200ms, you have severe bufferbloat. If your idle ping is 150ms (satellite), 200ms under load is fine. By measuring idle first, the A–F grade reflects your router’s behavior, not your ISP’s distance to the server.
Phase 2 — Download load test
A 25 MB file is downloaded via XHR while latency probes fire every second for 8 seconds. XHR’s onprogress event gives real-time byte counts so download speed updates live. The latency chart updates with each probe, color-coded green/amber/red vs your baseline so you can see exactly when bufferbloat begins to bite.
Phase 3 — Upload load test
2 parallel POST streams upload 1 MB chunks continuously for 8 seconds while probes measure latency. Upload bufferbloat is tested separately because cable and DSL connections have asymmetric bandwidth — the upload channel is narrower and fills much faster, often producing worse bufferbloat grades than download. Upload bufferbloat is the primary cause of lag during video calls.
| Tool | Idle Baseline | Download Test | Upload Test | A–F Grade | Browser |
|---|---|---|---|---|---|
| This test | ✓ Yes | ✓ 25 MB | ✓ Yes | ✓ Yes | ✓ Yes |
| Waveform Bufferbloat Test | ✓ Yes | ✓ Yes | ✓ Yes | ✓ Yes | ✓ Yes |
| DSLReports Speed Test | ✓ Yes | ✓ Yes | Partial | ✓ Yes | ✖ Defunct |
| Cloudflare Speed Test | ✓ Yes | ✓ Yes | ✓ Yes | ✖ No | Separate site |
| Speedtest.net | ✖ No | ✓ Yes | ✓ Yes | ✖ No | ✓ Yes |
How is this different from Waveform’s test? Waveform’s bufferbloat test uses WebSockets and their own test server infrastructure. This tool uses standard XHR requests to Cloudflare’s publicly available speed test endpoints — the same infrastructure used by millions of devices. The test methodology (baseline → download load → upload load) is similar, but this runs directly embedded in the page with no redirects or iframes.
Why does this matter more than a regular speed test? A connection that grades D or F on this test will cause serious problems for gaming and video calls even if your speed test shows 300 Mbps. Bufferbloat is the hidden reason many people with “fast” internet connections still experience choppy calls and game lag — and it’s fixable with the right router settings.
Bufferbloat Questions
How this works: Phase 1 sends 10 idle latency probes to speed.cloudflare.com via XHR. Phase 2 downloads a 25 MB file with XHR while probing latency every second for 8 seconds. Phase 3 POSTs 8 sequential 1 MB chunks while probing latency for 8 seconds. The worst latency increase vs baseline across both phases determines the A–F grade. No plugins or apps required — runs entirely in your browser.