|
|
|
|
|
by ar0
3588 days ago
|
|
While this is true for a high availability data center whose customers expect their website to be up 100% of the time, this article talks about storage infrastructure for their archiving service (C14). Restoring data from C14 will take multiple hours anyways, so networking downtime (or even power downtime, assuming they have enough backup power to safely power down their disks) will be of much lesser concern in the given use case than physical protection from natural or man-made disasters, which would destroy the archived data, or theft. The main argument I would agree with is that it would maybe be cheaper (and definitely more secure) to keep data redundantly in two data centers sufficiently spread out to not be affected by a singular disaster than in a single hyper-secure data center. |
|