3 ms·

Oh... ASMedia... Judging by their older controllers, they're going to need a heatsink, and even then they can fail out of the blue... At least it's SATA. This

by bserge 6y ago

Oh... ASMedia... Judging by their older controllers, they're going to need a heatsink, and even then they can fail out of the blue... At least it's SATA. This is exactly what I've been looking for, a cheap Ethernet connected device that can saturate 1GbE. It's perfect for backups.

rektide 6y ago
Well, cheap may rule this out... for now, but... Skip the middleman and directly connect your hard drive[1]! I think you may have trouble saturating the hard drive's dual 25Gbe links, and a decent switch is gonna be pricey, but that backup should be going very VERY fast indeed when you're done. Kioxia (nee Toshiba) is not the only one doing this, thankfully, cause I think it's interesting as heck. A lot of hard drives have a good amount of speedy Cortex R processors on them already, why use those to talk NVMe or SATA when you could have the hard drive talking ethernet directly? [1] https://business.kioxia.com/en-us/news/2020/ssd-20200922-2.html https://business.kioxia.com/en-us/news/2020/ssd-20200922-2.h...
- JAlexoid 6y ago
  When you're looking for durable long term backup solution - it's either high TBW values for NVMe drives or, practically, hard drives. Paying about $1000 for 20TB of fault tolerant extra long term storage isn't a problem. I'm even starting to consider magnetic tapes for archives. What's the price of a TB of high TBW NVMe storage?
  - rektide 6y ago
    I really wish we had better understanding about the failure modes of NVMe past wearing out the flash. Lets say we're writing a vibration sensor's data to the drive, once an hour. We leave 50% of the drive unpartitioned to give the drive controller plenty of headroom for whatever it wants. It might take centuries to get to the expected TBW. Yet there is still some kind of MTBF that is likely to happen. This above workload scenario is 100% write based. What about drives in data-centers that have lots of user images that they are unlikely/not-incentivized to prune. The drive could, hypothetically, only need to be written once, then forever read. When do sectors start going bad here? Or is the failure mode otherwise, in other componenets? Will this drive last a century? More? What other modes of failure should we expect, how probable are they? I love your challenge here @JAlexoid. Switching from thinking in terms of TB/$ to TB/y/$ is a great twist.

by bserge 6y ago

rektide 6y ago

Well, cheap may rule this out... for now, but... Skip the middleman and directly connect your hard drive[1]! I think you may have trouble saturating the hard drive's dual 25Gbe links, and a decent switch is gonna be pricey, but that backup should be going very VERY fast indeed when you're done. Kioxia (nee Toshiba) is not the only one doing this, thankfully, cause I think it's interesting as heck. A lot of hard drives have a good amount of speedy Cortex R processors on them already, why use those to talk NVMe or SATA when you could have the hard drive talking ethernet directly? [1] https://business.kioxia.com/en-us/news/2020/ssd-20200922-2.html https://business.kioxia.com/en-us/news/2020/ssd-20200922-2.h...

JAlexoid 6y ago
When you're looking for durable long term backup solution - it's either high TBW values for NVMe drives or, practically, hard drives. Paying about $1000 for 20TB of fault tolerant extra long term storage isn't a problem. I'm even starting to consider magnetic tapes for archives. What's the price of a TB of high TBW NVMe storage?
- rektide 6y ago
  I really wish we had better understanding about the failure modes of NVMe past wearing out the flash. Lets say we're writing a vibration sensor's data to the drive, once an hour. We leave 50% of the drive unpartitioned to give the drive controller plenty of headroom for whatever it wants. It might take centuries to get to the expected TBW. Yet there is still some kind of MTBF that is likely to happen. This above workload scenario is 100% write based. What about drives in data-centers that have lots of user images that they are unlikely/not-incentivized to prune. The drive could, hypothetically, only need to be written once, then forever read. When do sectors start going bad here? Or is the failure mode otherwise, in other componenets? Will this drive last a century? More? What other modes of failure should we expect, how probable are they? I love your challenge here @JAlexoid. Switching from thinking in terms of TB/$ to TB/y/$ is a great twist.