Hacker News Reader

Comment by pjc50 | Hacker News Reader

So how is this implemented? Spun-down disk drives?

My guess would be disk drives that are otherwise used for I/O intensive purposes, and where there's tons of unused/wasted space.

4MB/s is only around 4% of the bandwidth of a modern 1TB hard disk.

Why would they want to risk the performance of the I/O intensive disks?

With a 3s average response time and presumably replication across multiple disks, you've plenty of scope to schedule requests in a way that they won't affect performance.

In 3s on one disk you can do roughly 5k seeks and read up to 300MB. They only need to do 1 seek and read 4MB.

You can store 1 Pb and get 4000 MB/s

And at 3MB/s per TB it will always take over 3 days to retrieve all of your data. The 2-5 second retrieval latency is irrelevant when you're going to be waiting for 3 days...

FWIW it's nominally 4MB/s/TB:

> You should expect 4 MB/s of throughput per TB of data stored as Nearline Storage.

But still, you are correct in that it will take about 1TB / 4MB/s = 2.9d to retrieve all your data.

If you need the ability to do a restore faster than this, then you need to pay more for storage, is all. For many of us, waiting 3d to recover from a catastrophic failure isn't a big deal.

Is 3 days particularly slow for restoring a petabyte of data from an offsite backup?

s/over 3 days/no more than 3 days/

Presumably distributed across 1000 times more disks.

In my experience, spun-down disk drives take in the ballpark of 10-15 seconds to spin back up.

Reply on news.ycombinator.com