[Cerowrt-devel] Cira.ca network measurement system

[Cerowrt-devel] Cira.ca network measurement system

[Rich Brown]

Last Saturday, I heard a piece on CBC Radio's Spark program about the CIRA in Canada and their effort to map out network performance in order to get better Internet service for everyone across Canada. You can listen to the piece at: http://www.cbc.ca/radio/spark/286-empathy-games-intangible-art-and-more-1.3073000/test-your-internet-performance-crowdsource-too-1.3073048

I went to their site http://cira.ca/performance and it seemed pretty good. There were a *ton* of juicy measurements from the TCP stack when you click the Advance results button. But - you guessed it - no latency measurements. 

So I sent them a note and within an hour (on a Saturday!) I got a note back from a real human. (Apparently, they've just rolled out the service, so they're working hard to be responsive to all comers.)

I gave them the whole story about bufferbloat, and how valuable it would be to have those latency measurements. The initial reaction (on a Saturday) was sincere interest but no promises because they were just rolling out the service. 

It turns out that CIRA is using M-LAB for their test infrastructure. When I first looked, that project seemed dormant on Google Code - not updated recently. However, they just moved NDT to github (https://github.com/ndt-project/ndt) a couple weeks ago. And on http://www.measurementlab.net/tools/ndt they mention a websockets (non-Java) version of the web test page. So perhaps there's hope for yet another good test suite. 

I sent my note to the generic IPT@cira.ca. If a couple more people sent notes of encouragement, perhaps they'll move it up in their queue.

And does anyone know about Measurement Labs? How might we cajole them into adding a bufferbloat/latency test?

Rich

Re: [Cerowrt-devel] [Bloat] Cira.ca network measurement system

[Eggert, Lars]

On 2015-5-18, at 23:48, Rich Brown <richb.hanover@gmail.com> wrote:
> 
> Last Saturday, I heard a piece on CBC Radio's Spark program about the CIRA in Canada and their effort to map out network performance in order to get better Internet service for everyone across Canada. You can listen to the piece at: http://www.cbc.ca/radio/spark/286-empathy-games-intangible-art-and-more-1.3073000/test-your-internet-performance-crowdsource-too-1.3073048

Similar service out of Finland: http://www.netradar.org/

Lars

> 
> I went to their site http://cira.ca/performance and it seemed pretty good. There were a *ton* of juicy measurements from the TCP stack when you click the Advance results button. But - you guessed it - no latency measurements. 
> 
> So I sent them a note and within an hour (on a Saturday!) I got a note back from a real human. (Apparently, they've just rolled out the service, so they're working hard to be responsive to all comers.)
> 
> I gave them the whole story about bufferbloat, and how valuable it would be to have those latency measurements. The initial reaction (on a Saturday) was sincere interest but no promises because they were just rolling out the service. 
> 
> It turns out that CIRA is using M-LAB for their test infrastructure. When I first looked, that project seemed dormant on Google Code - not updated recently. However, they just moved NDT to github (https://github.com/ndt-project/ndt) a couple weeks ago. And on http://www.measurementlab.net/tools/ndt they mention a websockets (non-Java) version of the web test page. So perhaps there's hope for yet another good test suite. 
> 
> I sent my note to the generic IPT@cira.ca. If a couple more people sent notes of encouragement, perhaps they'll move it up in their queue.
> 
> And does anyone know about Measurement Labs? How might we cajole them into adding a bufferbloat/latency test?
> 
> Rich
> _______________________________________________
> Bloat mailing list
> Bloat@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/bloat