Hello, at the moment I'm seeing a lot of measurements on my both probes, that doesn't have any live target and, sometimes, have even invalid targets like 2001:db8:0:dead:beef::1 which is part of the IPv6 for documentation purposes. I've made a screenshot here: http://i.imgur.com/PRXHsOZ.png Can someone from RIPE check this out? Not sure if there is someone abusing RIPE Atlas in a nasty way. Michael
Hi Michael, On 2013/08/26 1:28 , Michael Stevens wrote:
at the moment I'm seeing a lot of measurements on my both probes, that doesn't have any live target and, sometimes, have even invalid targets like 2001:db8:0:dead:beef::1 which is part of the IPv6 for documentation purposes. I've made a screenshot here: http://i.imgur.com/PRXHsOZ.png
Can someone from RIPE check this out? Not sure if there is someone abusing RIPE Atlas in a nasty way.
I have no idea what the purpose is of those measurements. But user that created them does get charged for the results. And probes return results. So from that point of view nothing is wrong. Philip
On Mon, Aug 26, 2013 at 02:37:34AM +0200, Philip Homburg wrote:
I have no idea what the purpose is of those measurements. But user that created them does get charged for the results. And probes return results. So from that point of view nothing is wrong.
maybe it's time to compile a filter list to counter reconnaissance of thge probes' network vicinity, which I don't think was the stated purpose of the Atlas project. -Peter
Dear all, The culprit for the high number of tests running on the RIPE Atlas platform would be me. I do apologize for not sending a notification prior to deploying the tests, we do understand that we are looking at quite a massive number of measurements. We are currently trying to slow down the rhythm of the traceroutes, in an effort to not be so aggressive. The purpose of the measurements is to establish the reachability of the limited-visibility IPv6 prefixes we were able to identify using the BGP Visibility Scanner. The measurements are a part of our efforts of enhancing our understanding on the results of the visibility tool for IPv6 prefixes. The RIPE Atlas team was kind enough to help us with our research and allowed us to use the Atlas platform. For more information about the BGP Visibility project, I kindly refer you to http://visibility.it.uc3m.es/ We've previously focused on the visibility of the IPv4 prefixes, please find a more detailed view on our results in the following RIPE Labs article: https://labs.ripe.net/Members/andra_lutu/the-bgp-visibility-scanner For any further inquiries, do not hesitate to contact us! I do apologize for any distress this might have caused! For the future, we will try to notify before deploying another large number of tests! Thank you, best regards, Andra
Hi, We will soon be resuming our measurements using the RIPE Atlas platform. We've scheduled the traceroutes to one every 18 seconds (a more uniform distribution than the previous approach of batches of 100 tests every 30 minutes), to address the concerns regarding the load we put on the system. We still have about 1,300 target addresses to test, so the total duration of the measurements is estimated to approx. 7 hours (i.e., scheduled to start at 16h00 CET -- with expected finishing time at around 23h00 CET). We do apologize for any inconvenience! For any further comments, inquiries, suggestions etc., do not hesitate to contact us! Thank you, best regards, Andra On 08/26/2013 10:23 AM, Andra Lutu wrote:
Dear all,
The culprit for the high number of tests running on the RIPE Atlas platform would be me. I do apologize for not sending a notification prior to deploying the tests, we do understand that we are looking at quite a massive number of measurements. We are currently trying to slow down the rhythm of the traceroutes, in an effort to not be so aggressive.
The purpose of the measurements is to establish the reachability of the limited-visibility IPv6 prefixes we were able to identify using the BGP Visibility Scanner. The measurements are a part of our efforts of enhancing our understanding on the results of the visibility tool for IPv6 prefixes. The RIPE Atlas team was kind enough to help us with our research and allowed us to use the Atlas platform. For more information about the BGP Visibility project, I kindly refer you to http://visibility.it.uc3m.es/ We've previously focused on the visibility of the IPv4 prefixes, please find a more detailed view on our results in the following RIPE Labs article: https://labs.ripe.net/Members/andra_lutu/the-bgp-visibility-scanner
For any further inquiries, do not hesitate to contact us! I do apologize for any distress this might have caused! For the future, we will try to notify before deploying another large number of tests!
Thank you, best regards, Andra
Hello, On 2013.08.26. 8:52, Peter Koch wrote:
On Mon, Aug 26, 2013 at 02:37:34AM +0200, Philip Homburg wrote:
I have no idea what the purpose is of those measurements. But user that created them does get charged for the results. And probes return results. So from that point of view nothing is wrong.
maybe it's time to compile a filter list to counter reconnaissance of thge probes' network vicinity, which I don't think was the stated purpose of the Atlas project.
-Peter
Indeed, we do exclude the "probes' network vicinity", as you called it. At the moment, this is the exclusion list: exclude4 = ('PRIVATE', 'RESERVED') exclude6 = ('LINKLOCAL', 'UNSPECIFIED', 'LOOPBACK', 'MULTICAST') (The above acronyms are defined in the IPy Python module, see https://github.com/haypo/python-ipy/blob/master/IPy.py for details.) Please speak up if you have justification for other networks to be excluded. Note that it's very difficult to apply the above filters to perfection; a sophisticated enough user may be able to find ways around it. Retards, Robert
participants (5)
-
Andra Lutu -
Michael Stevens -
Peter Koch -
Philip Homburg -
Robert Kisteleki