3 ms·
I think he's referring to AWS which sites their datacenters and have multiple fiber connections situated such that cutting one doesn't bring down the entire dat
by rstupek 11y ago
I think he's referring to AWS which sites their datacenters and have multiple fiber connections situated such that cutting one doesn't bring down the entire datacenter which is what appears to have happened here. Running multiple fibers likely costs more, ergo a place where OVH cut features (and redundancy) to have a lower price
- joepie91_ 11y agoAt OVH's scale, I doubt that this is a realistic cost issue. In fact, it is my understanding that OVH was already planning to bring more fiber connections online, and they've accelerated that process in order to get their network back up today. I don't know. It just irks me when people start making these kind of statements when they almost certainly do not have the data needed to support them. It smells of hype to me, especially where it concerns the 'hip' services like AWS, DigitalOcean, (previously) Linode, and so on. Trying to find some way to turn the narrative into something that puts those providers in a positive light, when it most likely had absolutely nothing to do with it. EDIT: And again, "it's cheaper so that must be why it has worse redundancy" is far, far too simplistic. There are many more variables at play.
- rstupek 11y agoWorse redundancy = cheaper is what I said, not the reverse. Amazon doesn't setup an AWS datacenter without multiple physically separated connections. The data to support my statement is in the fact they had 1 line connecting the data center and it was cut taking everything out. They may have started making additional lines into the center but they didn't have them today.