Skip to Content.

rare-dev - Re: [rare-dev] perfSONAR on RARE/FreeRtr

Subject: Rare project developers

List archive


Re: [rare-dev] perfSONAR on RARE/FreeRtr


Chronological Thread 
  • From: Frédéric LOUI <>
  • To: Tim Chown <>
  • Cc: "" <>, Xavier Jeannin <>, Tomasz Szewczyk <>, Ivana Golub <>
  • Subject: Re: [rare-dev] perfSONAR on RARE/FreeRtr
  • Date: Thu, 25 Aug 2022 13:13:20 +0200
  • Dkim-filter: OpenDKIM Filter v2.10.3 zmtaauth04.partage.renater.fr 9BE5A1C00F5

Thanks Tim for your explanation.

I’m not super familiar with Lola but do you think that would be a relevant
demo show casing:

- Lola streaming UDP to RARE/freeRtr unicat -> multicast translator
- inject this stream into GP4L BIER multicast network
- An a JunoS AMT relay in GP4L show casing interoperability
- we can still have native multicast reception in GP4L (my homenet is
configured for that for example ...)
- we also can have the multicast menu updated automatically so people can see
when a Lola streaming is effective

FYI the stream can be any where in Europe, we can receive can this multicast
stream via BIER all the way to CALTECH’s TOFINO2 starting from today ;)

Then this stream can be monitored in real time via Timemap fed with GP4L
either via streaming telemetry or prometheus

Please let me know your (or anyone’s) thoughts

All the best
Frederic

> Le 25 août 2022 à 11:57, Tim Chown <> a écrit :
>
>> On 25 Aug 2022, at 10:41, mc36 <> wrote:
>>
>> On 8/25/22 11:34, mc36 wrote:
>>> then 2 constant rsvp tunnels and that's it...
>>
>> first you must nat the unicast to multicast... rare and cat6500 both can
>> do that in hardware... :)
>> then you static joingroup the resulting stream to the 2 tunnel-te s
>> and that's all... you have 2 constant streams, one copy should reach the
>> tunnel destination...
>> if not, you can have more constant streams... no need to do a thing in the
>> core if there are some bandwidth...
>>
>> so which netmusic.... the distributed musicinas (lola?:) project you
>> mentioned yesterday…
>
> Aaaah! :)
>
> So that type of distributed music performance uses platforms like LoLa,
> which is a project led by GARR in Italy - https://lola.conts.it/, and the
> work in WP6T1 on LoLa is really about measuring latency and jitter to help
> troubleshoot networks when the distributed musicians get poor quality of
> experience. That’s what’s led to the TimeMap platform that gives a view of
> the latency and jitter per hop on the GEANT backbone.
>
> LoLa, the software, I think is Windows only, and you run it on very well
> tuned hardware (as encoding and decoding adds latency, maybe more than the
> network). It doesn’t use RTP, just raw UDP, as far as I know.
>
> There’s a map of known LoLa users here, seems to be one in Budapest too!
> https://www.google.com/maps/d/u/0/viewer?mid=1RUFFnVkuPnXnBPWThnUlMp96_Co&hl=it&ll=-3.81666561775622e-14%2C-82.15722249999999&z=1
>
> The interesting thing is then whether TimeMap can be easily applied to the
> GEABT P4 Lab network. It seems it can. There are of course other tools
> you could use, but validating TimeMap on our P4 Lab would be useful, if
> it’s not a lot of effort. It all seems compatible with Telegraf export to
> influxdb.
>
> Tim
>




Archive powered by MHonArc 2.6.19.

Top of Page