[Bloat] Comparing bufferbloat tests (was: We built a new bufferbloat test and keen for feedback)

[Arshan Khanifar]

Hello Toke!

> Curious to know why you picked this instead of, say, something from
> speed.cloudflare.com (since you're using that for the speed tests anyway)?

The reason we chose the static file instead of cloudflare was because the ping to Cloudflare from where I am reported a pretty high number (around 20ms), and we also noticed that there was an issue with the network path to cloudflare in general. We were getting some packet losses when doing a normal ping, and there were some latency spikes that would happen regularly. 

> Yeah, allowing some ramp-up time before determining the bandwidth seems
> reasonable, but it's not generally possible to just pick a static number
> of (say) seconds to chop off... Having the graph over time helps
> sanity-check things, though.

Yeah it’d be useful to have a real-time graph. 

> Also, a continuous graph of latency samples over time (for the whole
> duration, including idle/upload/download) is usually very instructive
> when plotting such a test.

I had that in the earlier version of the test that I sent to you previously. It would make the test look a bit complicated though, so we should maybe have an advanced view, where the user can see these charts.

> The test at speed.cloudflare.com does seem to plot real-time upload
> bandwidth; is that a privileged operation for themselves, or something?

It’s not “that” real time, they start from a small upload size, and progressively increase the file size. They will calculate the average bandwidth for each upload and show it on the chart that they have. I want to have the test continuously do a big upload, and have real-time measurements.

> If you're reporting standard deviation, I'd just label it as such. One
> good definition for jitter is IPDV:
> https://en.wikipedia.org/wiki/Packet_delay_variation

Thanks! This is great! 

> 
>> @Toke Høiland-Jørgensen
>>> Also, what are the shields below the grade supposed to mean? Do they
>>> change depending on the result? On which criteria? 
> 
> Right, cool; explaining this on the page might be useful ;)

Yeah we’ll change the texts and add more explanations in the end. 

> Just having the test be on a separate (static) page would be the obvious
> fix ;)

Correct hahah. I’ll bring that up with the team. 

Thank you so much for all the help!

Arshan