Skip to content

Add Bufferbloat.org to web-based bufferbloat tests - #95

Draft
pelagus wants to merge 1 commit into
tohojo:masterfrom
pelagus:add-bufferbloat-org-test-links
Draft

Add Bufferbloat.org to web-based bufferbloat tests#95
pelagus wants to merge 1 commit into
tohojo:masterfrom
pelagus:add-bufferbloat-org-test-links

Conversation

@pelagus

@pelagus pelagus commented Jul 24, 2026

Copy link
Copy Markdown

Thank you for maintaining Bufferbloat.net and the body of documentation around bufferbloat, CoDel, FQ-CoDel, CAKE, and practical testing. I maintain Bufferbloat.org, and I treat this work as foundational: the Bufferbloat.org learning bibliography explicitly describes these sources as "sources I respect and consider foundational to this work", with links to the Bufferbloat Project wiki, Bufferbloat.net, and the CAKE documentation.

This pull request adds Bufferbloat.org's browser-based internet reliability test to existing Bufferbloat.net pages that already point readers to web-based bufferbloat tests. The intent is to add one more test option for readers who are trying to measure latency under load, while keeping the existing page structure and recommendations intact, without implying any official affiliation or endorsement.

The new link is placed immediately after the Waveform test where that ordering exists, and uses the anchor "Bufferbloat.org Internet Reliability Test" where the surrounding page style supports a descriptive label. The main Tests for Bufferbloat page also adds a short description:

Browser-based test of latency during download and upload load, with open-source implementation and published methodology.

The linked test page is:

https://bufferbloat.org/test

The anchor text is lightly adapted from the page title, "Test Internet Reliability: Speed, Latency, and Bufferbloat", to fit the style of the existing lists.

Files changed:

  • content/bloat/wiki/Tests_for_Bufferbloat.md
  • content/bloat/wiki/What_can_I_do_about_Bufferbloat.md
  • content/bloat/wiki/index.md
  • content/bloat/wiki/TechnicalIntro.md
  • content/bloat/wiki/Bufferbloat_FAQs.md
  • content/cerowrt/wiki/index.md

Validation:

  • Ran git diff --check
  • Did not run Hugo preview because this environment does not have hugo or docker installed; the project README notes Hugo 0.16 is required for local preview.

@FrantisekBorsik

Copy link
Copy Markdown

I don’t like your demand to have this unknown test placed right under Waveform one. Doesn’t make sense. Even our test, LibreQoS Bufferbloat Speedtest, the best speedtest out there, was placed down there, basically in the stacking order of appearance on the market.

Your test need to be tested thoroughly, before being added.

Try to share it with the community at bufferbloat mailing lists first and get the feedback, then improve it.

Your starting point: https://lists.bufferbloat.net/

@richb-hanover

Copy link
Copy Markdown
Collaborator

First off, thanks for considering a contribution. I did try your tester, and it certainly is attractive and seemed OK in a one-out-of-one test. (I like how you forced the browser to full-screen - I suspect that keeps the test page in the foreground for more accurate measurements.)

But I second Frank's request: There are lots of "bufferbloat tests" out there. To be listed on the Bufferbloat.net page, I'd like to know answers to these questions:

  • What distinguishes bufferbloat.org from any of the others: LibreQoS test, speedtest.net, the Waveform test page, etc.?

  • What testing have you done to assert that your test results are accurate? Actually looking at tcpdump? Comparing results from multiple runs of bufferbloat.org against those others?

  • Finally, you should definitely open up testing to the wider community (starting with the Bloat mailing list at list.bufferbloat.net) to "get a bunch of miles on it".

Thanks again.

@pelagus

pelagus commented Jul 25, 2026

Copy link
Copy Markdown
Author

@FrantisekBorsik that's fair. I placed it next to Waveform because grouping browser tests together seemed clearer from a user view, but I see the preference for chronological order and I'm happy with whatever the maintainers decide.

@richb-hanover thanks for testing it, and glad you noticed the fullscreen. I was curious if anyone would pick up on that.

Unlike most of the bufferbloat tests out there, this one is fully open source with published methodology, and every ping sample is inspectable and exportable. The grade translates the measurements into what they mean for calls and games, rather than stopping at the numbers. I've tested it on fiber, cable, mobile, and VPN and the grading catches bufferbloat reliably when it's there. Haven't done formal side-by-side benchmarks yet and I should, and that's part of why I'm gladly heading to the mailing list now.

My hope is this complements the tools already out there rather than competes with them. I built it because I wanted something with an open methodology and user-education focused. Thanks for the thoughtful questions.

@FrantisekBorsik

FrantisekBorsik commented Jul 25, 2026

Copy link
Copy Markdown

Well, no @pelagus - LibreQoS is browser-based test as well, in fact, all of them are, even though some of them have also apps.

Do you rounds on the mailing lists, get your feedback, improve the test and then let's see. Being open source it's not a means to an end, or argument to get on the bufferbloat.net page per se, or whatever else. LibreQoS test was fully open source, but we were forced to Cloudflare it for now, we are considering to release it. Even CLI version was added for IETF 126 Vienna, last week.

To help you out, here are the lists that are relevant, so start with the introduction of your test there, please:

Another helpful thing for you would be to read Dave's blog post(s) on flaws of modern speedtests, maybe you will find something useful to implement in your test: https://blog.cerowrt.org/post/speedtests/

@pelagus

pelagus commented Jul 26, 2026

Copy link
Copy Markdown
Author

Oh, I see what happened now. There are several pages that mention bufferbloat tests in different contexts, and I ended up applying the same change to sections that explicitly list only web-based tests.

I'll move this back to draft and take the test through community review and validation first. If and when it feels ready, I'll come back with a more structured change proposal, possibly suggesting a single canonical page for bufferbloat tests to make maintenance easier in the future.

Thanks again for the guidance.

@pelagus
pelagus marked this pull request as draft July 26, 2026 11:15
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants