[Bloat] random window into other people's latency experience

[Bloat] random window into other people's latency experience

[Sebastian Moeller]

[-- Attachment #1: Type: text/plain, Size: 1914 bytes --]

Dear All,

I use free thinkbroadband's, latency/availability monitoring  (broadban quality monitoring BQM, free after a free registration*)service, which runs ICMP probes from a dedicated box in London IIRC against all registered IPv4-addresses/domain names. Yesterday, my dsl-modem and the DSLAM decided they do not like each other and the modem was in idle_request status for several hours from afternoon to evening. Interstingly the latency probe was continuing, but now not against may router, but against the german Telefonica/O2 customer who got my old dynamic IP address assigned:

<a title="Broadband Ping" href="https://www.thinkbroadband.com/broadband/monitoring/quality/share/889b0c5ebab7cd432400d55e6c414b6f6658e278-09-06-2020"><img alt="My Broadband Ping - sesaokjole" src="https://www.thinkbroadband.com/broadband/monitoring/quality/share/889b0c5ebab7cd432400d55e6c414b6f6658e278-09-06-2020.png" /></a>

As a belts AND suspenders measure I also try to attach an image, but  am not sure whether the list will like that...



From 12:30 to ~21:00 on the plot my modem was out of sync and hence these probes must have been reflected from somebody else's router in Germany in the same ISP's network as my link. As the green, blue and yellow lines indicate that access link suffers much worse from variable latency (and from 19:30 - 21:00 also from some packet loss). I believe I have forgotten how bad internet access can be without competent AQM to make performance under load acceptable.

So here is to all who worked hard to diagnose, understand and to remedy bufferbloat.

Best Regards
	Sebastian


*) This is similat to dslreports smokeping service, except it is closer for european users and it allows domain names instead of fixed IP-addresses and will by default run wiyhout a time limit, the reporting is not half as nice as with dslreports, but for free it is quite marvelous.

[-- Attachment #2.1: Type: text/html, Size: 3051 bytes --]

[-- Attachment #2.2: 889b0c5ebab7cd432400d55e6c414b6f6658e278-09-06-2020.png --]
[-- Type: image/png, Size: 239410 bytes --]

Re: [Bloat] random window into other people's latency experience

[Jonathan Foulkes]

[-- Attachment #1: Type: text/plain, Size: 2648 bytes --]

Always fun to see ping-porn, I have a huge trove myself ;-)

And same, a big thanks to everyone, including you Seb.

Best regards,

Jonathan Foulkes


> On Jun 10, 2020, at 6:01 AM, Sebastian Moeller <moeller0@gmx.de> wrote:
> 
> Dear All,
> 
> I use free thinkbroadband's, latency/availability monitoring  (broadban quality monitoring BQM, free after a free registration*)service, which runs ICMP probes from a dedicated box in London IIRC against all registered IPv4-addresses/domain names. Yesterday, my dsl-modem and the DSLAM decided they do not like each other and the modem was in idle_request status for several hours from afternoon to evening. Interstingly the latency probe was continuing, but now not against may router, but against the german Telefonica/O2 customer who got my old dynamic IP address assigned:
> 
> <a title="Broadband Ping" href="https://www.thinkbroadband.com/broadband/monitoring/quality/share/889b0c5ebab7cd432400d55e6c414b6f6658e278-09-06-2020 <https://www.thinkbroadband.com/broadband/monitoring/quality/share/889b0c5ebab7cd432400d55e6c414b6f6658e278-09-06-2020>"><img alt="My Broadband Ping - sesaokjole" src="https://www.thinkbroadband.com/broadband/monitoring/quality/share/889b0c5ebab7cd432400d55e6c414b6f6658e278-09-06-2020.png <https://www.thinkbroadband.com/broadband/monitoring/quality/share/889b0c5ebab7cd432400d55e6c414b6f6658e278-09-06-2020.png>" /></a>
> 
> As a belts AND suspenders measure I also try to attach an image, but  am not sure whether the list will like that...
> <889b0c5ebab7cd432400d55e6c414b6f6658e278-09-06-2020.png>
> 
> 
> From 12:30 to ~21:00 on the plot my modem was out of sync and hence these probes must have been reflected from somebody else's router in Germany in the same ISP's network as my link. As the green, blue and yellow lines indicate that access link suffers much worse from variable latency (and from 19:30 - 21:00 also from some packet loss). I believe I have forgotten how bad internet access can be without competent AQM to make performance under load acceptable.
> 
> So here is to all who worked hard to diagnose, understand and to remedy bufferbloat.
> 
> Best Regards
> 	Sebastian
> 
> 
> *) This is similat to dslreports smokeping service, except it is closer for european users and it allows domain names instead of fixed IP-addresses and will by default run wiyhout a time limit, the reporting is not half as nice as with dslreports, but for free it is quite marvelous.
> _______________________________________________
> Bloat mailing list
> Bloat@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/bloat


[-- Attachment #2: Type: text/html, Size: 4111 bytes --]

Re: [Bloat] random window into other people's latency experience

[Chris Wedgwood]

the graph/bitmap it creates is actually quite terse but also useful at
the same time

how often does it probe?  (i'm curious about the number of data-points
per pixel)

Re: [Bloat] random window into other people's latency experience

[Sebastian Moeller]

Hi Chris,

some details below.

> On Jun 12, 2020, at 00:57, Chris Wedgwood <cw@f00f.org> wrote:
> 
> the graph/bitmap it creates is actually quite terse but also useful at
> the same time

	It is like smokeping for a non-tech audience...

> 
> how often does it probe?  (i'm curious about the number of data-points
> per pixel)

According to https://www.thinkbroadband.com/faq/broadband-quality-monitor they send an ICMP echo request (aka ping) once every second and each horizontal pixel seems to aggregate 100 seconds.

Best Regards
	Sebastian