Bufferbloat Test (Latency Under Load)
Compare your connection response time when it is idle and while it is busy downloading and uploading, to see whether heavy transfers make calls and games lag.
Response time of repeated HTTPS requests to this site while the connection is idle, then under a short, capped download load, then under a short, capped upload load. Close other downloads and streaming, keep the tab visible, read the data notice and press Start. The test takes about 30 seconds and uses up to about 60 MB of download and 12 MB of upload.
What does this test measure?
Response time of repeated HTTPS requests to this site while the connection is idle, then under a short, capped download load, then under a short, capped upload load.
How do you run the test?
Close other downloads and streaming, keep the tab visible, read the data notice and press Start. The test takes about 30 seconds and uses up to about 60 MB of download and 12 MB of upload.
How should results be interpreted?
A small increase under load is good. A large increase means the connection becomes sluggish when it is busy, which can cause lag in calls and games while someone downloads or uploads.
Technical limitations
This is not ICMP ping and not a packet-loss measurement. The browser, this server and other devices on the network influence the result, and a very fast connection may finish the capped load too quickly for a reliable result.
Why can the score decrease?
Unstable measurements, denied permissions, unsupported browser APIs, background load, power saving, or an incomplete test can affect the result. When a score changes, the result screen explains the reason.
What should you do if a problem is found?
Try Ethernet instead of Wi-Fi, turn on smart queue management (SQM/QoS) on a router that supports it, limit heavy transfers during calls, and test again.
Example Result Explanation
Frequently Asked Questions
Does this test provide a definitive hardware diagnosis?
No. The test measures signals the browser can safely access and clearly explains its technical limits.
Why can results change between runs?
Background applications, power mode, network conditions, permissions, and operating-system timing can affect results.
How should I repeat the test if I see a problem?
Repeat it on the same device and browser while changing only one condition, so the source of the difference is easier to identify.