As the discussion rages on NANOG, RIPE, CENTR and many other uber-technical forums, I would like to see whether we can focus on what we know best - networking. Perhaps a weekly report of fiber cuts throughout Europe (starting from Feb 15) and the RFO that the carrier provided. Of especial interest would be undersea/underocean cuts or strange outages that the carrier cannot explain. Perhaps we can then map where some nation/state is sabotaging fiber or tapping into such fiber.
Anyone willing to run with this?
Infrapedia seems like a logical place to aggregate such data.
At least 40 years of telco's being unwilling to share information about their service interuptions? They are barely willing to acknowledge outages to their own customers.
You do realise there's a shedload of fibre running around Europe ? There are so many redundant paths that you'd have to chop through quite a lot of it before anyone noticed much difference.
I mean even within Ukraine itself, traditional internet (i.e. non-satcom) has proven to be surprisingly resilient.
Historically, the largest telecommunication outages have been due to operator error & software bugs (or the malicious equivalent). Problems in large networks have more impact.
Vulnerability to physical damage varies a lot between countries. Some countries have a lot of redundancy, other countries have limited redundancy.
The big question for decision makers are trends.
Are these 'normal' outage trends or 'unusual' outage trends. One farmer on a tractor digging up fiber may be normal. Two farmers is a concidence. More than two farmers is abnormal.
<Insert my usual the importance of information sharing speech here>
Yeah, tell me about it.
In the very recent past there was a certain Tier 1 operator who decided to move a core router between X & Y within the same datacentre. As you might imagine, this impacted a good few hundred fibres.
Physical move happened without issue, router powered up fine, but a large subset of fibres refused to come up.
Cue escalation, lots of testing, lots of talking to vendors etc.
Eventually 72 hours later the problem was solved.
Cause ? SFP signals too high at new location, attenuators fitted, everything came up.