Does the Internet Route Around Damage? - Edition 2023
• 5 min read
On 22-23 November 2023, the Amsterdam Internet Exchange (AMS-IX) had an outage on their Amsterdam peering platform. As this is generally considered one of the largest IXPs on the planet, we looked into this event from the perspective of RIPE Atlas to see what it had to tell us about the robustness …
“Ghost routes: https://www.sixxs.net/tools/grh/what/”
I've added a reference to the different names 'stuck routes' and 'ghost routes' for this phenomenon at the beginning of the post. Thanks for the pointer!
“Interested in repeating this analysis for 2018 world cup?”
Hi Dan, we have no plans of repeating this analysis this time. This type of signal is still there, see for instance https://twitter.com/search?q=%40ohohlfeld%20%23worldcup&src=typd for a couple of graphs that show the impact in various places.
“@emileaben Rather than standardizing human-readable output format, why not emitting a standard structured format, separating the network part (traceroute) and the visualisation part (a tool using the structured output format). Such a format already exists, in RFC 5388. I let you do the same in JSON :-)”
thanks for the interest in the topic Stephane. You hit the nail on the head, the main idea was to standardise a structured format for traceroute. I notice a lack of enthusiasm for RFC5388, probably due to it's verbosity. quick test shows that gzip compression of RFC5388-style results would need 3x more storage relative to plain-text traceroute results. But the RFC is likely very useful to see if we cover all bases in a slimmer structured output format.
One other activity that may be worth mentioning here: We organised a get-together for traceroute implementers. As many traceroute implementations do things slightly different, a bit more coordination can help in making things more consistent, for instance in output formats.
“I'm trying to work with the ixp-jedi tool. In this step: ## measure.py This script runs one-off measurements for the probes specified in _probeset.json_ and stores their results in _measurementset.json_ This uses the RIPE Atlas measurement API for measurement creation, And it needs a valid measurement creation API key in ~ / .atlas / auth When trying to execute the script ./measure.py I get the following and I do not know how to solve it. Authentication file /root/.atlas/auth not found Please, I need your help.”
hi, thanks for trying to use the tool. i hope the docs on github are clear enough: https://github.com/emileaben/ixp-country-jedi/#measurepy --- This script runs one-off measurements for the probes specified in probeset.json and stores their results in measurementset.json This uses the RIPE Atlas measurement API for measurement creation, and it needs a valid measurement creation API key in ~/.atlas/auth . For more information on RIPE Atlas API keys see https://atlas.ripe.net/docs/keys/ --- if not let me know how to improve that. if you are interested in country-level monthy runs. these are available at: http://sg-pub.ripe.net/emile/ixp-country-jedi/history/
“Hi, Is there a way to download multiple days dataset without having to do them individually? Also do you have any API's which will permit me to download the datasets using wget?”
Hi Meenakshi, I think you'll have to download the files individually. I think, if your RIPE Access account doesn't have 2 factor authentication, you can use wget to download the files with the --user and --password options.
While we were busy pushing this post out, it looks like the Gambian Internet was restored, roughly around 12h UTC on 2 December. RIPE Atlas probes came online again, and we see 6 out of 7 ASNs in RIS data again.
Showing 17 comment(s)