Updated RIPE Atlas roadmap
Dear colleagues, we have considered the feedback we have received from you in the previous few months, on this list and MAT-wg mailing list, and last week we have made an updated version of the roadmap: http://roadmap.ripe.net/ripe-atlas/ We have added few new items under "requested features", and shifted some of the previously requested and planned towards "in progress". We are updating this document every two months, so you can expect newer requests to be added early next year. Regards, Vesna
Hi, Why is 'quick Look DNS' not available for all users? Also, I'd love to have constrains for probe selection. This would allow to ping from i.e. 3 probes in each AS. Please consider such a feature. Thanks, Dave. ----- Quick Look DNS Introduce easy to use 'Quick Look' DNS measurements with rapid results for RIPE NCC members
Op 19 dec. 2013 om 16:44 heeft Vesna Manojlovic <BECHA@ripe.net> het volgende geschreven:
Dear colleagues,
we have considered the feedback we have received from you in the previous few months, on this list and MAT-wg mailing list, and last week we have made an updated version of the roadmap:
http://roadmap.ripe.net/ripe-atlas/
We have added few new items under "requested features", and shifted some of the previously requested and planned towards "in progress".
We are updating this document every two months, so you can expect newer requests to be added early next year.
Regards, Vesna
On 2013/12/19 19:26 , dave wrote:
Also, I'd love to have constrains for probe selection. This would allow to ping from i.e. 3 probes in each AS. Please consider such a feature.
Just thinking out loud, is this something that should be part of the web UI? Before you know it, probe selection becomes some sort of complex filter language that can do anything, but nobody knows how to use it. Instead, it may be worth doing these kinds of things through the API: download the complete list of probes, apply your filter criteria and then create a measurement that involves the probes that match. Philip
Hi Philip, Working via API is not something anyone can do. Also, it seem to me to be something for a measurement which was given quite some thought. The web UI allows novice users to do a quick test as a tool for faultfinding. No API knowledge needed. That is what I'm using Atlas for. Quick short tests from multiple points. Thanks for your response.
Op 19 dec. 2013 om 19:34 heeft Philip Homburg <philip.homburg@ripe.net> het volgende geschreven:
On 2013/12/19 19:26 , dave wrote: Also, I'd love to have constrains for probe selection. This would allow to ping from i.e. 3 probes in each AS. Please consider such a feature.
Just thinking out loud, is this something that should be part of the web UI? Before you know it, probe selection becomes some sort of complex filter language that can do anything, but nobody knows how to use it.
Instead, it may be worth doing these kinds of things through the API: download the complete list of probes, apply your filter criteria and then create a measurement that involves the probes that match.
On 2013/12/19 21:25 , dave wrote:
Working via API is not something anyone can do. Also, it seem to me to be something for a measurement which was given quite some thought.
The web UI allows novice users to do a quick test as a tool for faultfinding. No API knowledge needed. That is what I'm using Atlas for. Quick short tests from multiple points.
Using the API does not imply that you have to write the script yourself. There have been quite a few requests for probe selection features. I sort of wonder if those can be implemented without leading to complexity that nobody can understand anymore. Philip
Please tell me more. Is there a repository of scripts? I've looked, but not yet found such. If a script is available to configure pings from maximum 3 probes in each AS in a selectable country to a selectable target, I'm interested. Anyway, if such could be implemented in the web UI I'd prefer it. Thanks.
Op 20 dec. 2013 om 13:06 heeft Philip Homburg <philip.homburg@ripe.net> het volgende geschreven:
Using the API does not imply that you have to write the script yourself.
There have been quite a few requests for probe selection features. I sort of wonder if those can be implemented without leading to complexity that nobody can understand anymore.
Philip
On Sat, Dec 21, 2013 at 9:06 AM, dave <gboonie@gmail.com> wrote:
Please tell me more. Is there a repository of scripts? I've looked, but not yet found such.
Some code can be found at GitHub [1]. The repo hosts content contributed by members of the RIPE Atlas community. You can find code to create measurements from the CLI. I particularly recommend Stéphane Bortzmeyer's Tutorial files. Regards, [1] https://github.com/RIPE-Atlas-Community/ripe-atlas-community-contrib
If a script is available to configure pings from maximum 3 probes in each AS in a selectable country to a selectable target, I'm interested.
Anyway, if such could be implemented in the web UI I'd prefer it.
Thanks.
Op 20 dec. 2013 om 13:06 heeft Philip Homburg <philip.homburg@ripe.net> het volgende geschreven:
Using the API does not imply that you have to write the script yourself.
There have been quite a few requests for probe selection features. I sort of wonder if those can be implemented without leading to complexity that nobody can understand anymore.
Philip
On Sat, Dec 21, 2013 at 09:06:09AM +0100, dave <gboonie@gmail.com> wrote a message of 21 lines which said:
Please tell me more. Is there a repository of scripts? I've looked, but not yet found such.
There is apparently possible improvments on the web site... Iñigo Ortiz de Urbina gave you the URL but do note there is today no script to apply Philip Homburg's excellent idea (download the list of probes, then filter, then create the measurement). TODO :-)
If a script is available to configure pings from maximum 3 probes in each AS in a selectable country to a selectable target, I'm interested.
Not yet. Anyone brave enough to do it and submit it to the Github repository?
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On 23/12/2013 10:05, Stephane Bortzmeyer wrote:
On Sat, Dec 21, 2013 at 09:06:09AM +0100, dave <gboonie@gmail.com> wrote a message of 21 lines which said:
Please tell me more. Is there a repository of scripts? I've looked, but not yet found such.
There is apparently possible improvments on the web site... Iñigo Ortiz de Urbina gave you the URL but do note there is today no script to apply Philip Homburg's excellent idea (download the list of probes, then filter, then create the measurement). TODO :-)
If a script is available to configure pings from maximum 3 probes in each AS in a selectable country to a selectable target, I'm interested.
Not yet. Anyone brave enough to do it and submit it to the Github repository?
Hi, I did have a similar itch-to-scratch, so I do have a hack laying around that almost does what you want, it's usage description: schootpunt$./select_probes.py -h usage: select_probes.py [-h] [-v] [-l LOCSTRING] [-p POINT] [-r RADIUS] [-c COUNT] [-f FIELDS] [-a MAXPERAS] [-d] optional arguments: -h, --help show this help message and exit -v, --verbosity -l LOCSTRING, --locstring LOCSTRING location string, ie 'Amsterdam,NL' -p POINT, --point POINT location as <lat>,<lon>-string, ie '48.45,9.16' -r RADIUS, --radius RADIUS radius (km) within which to select probes -c COUNT, --count COUNT number of probes requested -f FIELDS, --fields FIELDS comma separated list of fields to return (output is tsv) -a MAXPERAS, --maxperas MAXPERAS maximum no. of probes per IPv4 source AS -d, --includedownprobes include probes that are down (default: don't include these) This doesn't select an individual country, but selects probes around a specific coordinate, so ./select_probes.py -l"Amsterdam,NL" -c 50 -a 3 -f id Would return the IDs of the 50 closest probes to Amsterdam city center (or whatever coordinate the google geocoding API gives for "Amsterdam,NL"), with a maximum of 3 probes per v4 AS. I can submit this to the Github, if it's close enough to what others would want. Emile -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.15 (Darwin) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iF4EAREIAAYFAlLB4TsACgkQj05ACITZaqopWAD9Fa3ri/TP05qNDQ7agglxVqCE nldKJEPLQ0PmPhDtPQ0BAIdGX6XagIjeBokd5LNiogCpra6yOOKnrrfqKfq4B6Y7 =TCFU -----END PGP SIGNATURE-----
On Fri, Dec 20, 2013 at 01:06:59PM +0100, Philip Homburg <philip.homburg@ripe.net> wrote a message of 12 lines which said:
There have been quite a few requests for probe selection features. I sort of wonder if those can be implemented without leading to complexity that nobody can understand anymore.
Seems reasonable. But, in that case, I would request, in exchange, that the default selection is improved to be actually random: there are apparently huge biases in the selection, so drawing quantitative conclusions ("x % of Atlas probes can ping www.example.com with IPv4") from default selection is dangerous.
participants (6)
-
dave -
Emile Aben -
Iñigo Ortiz de Urbina -
Philip Homburg -
Stephane Bortzmeyer -
Vesna Manojlovic