If Google can build the worlds largest clustered computing environment out of garbage, you can build enough redundancy into hardware this cheap to have your cake and not pay as much for it.
You could even mirror these appliances for total redundancy way less than the cost of most large tier storage solutions.
Besides, just because something says "HP" doesn't mean it's not garbage. It often is; it's just well supported garbage.
lol you can build the same if you use 666 1.5 Terabyte drives at $97.99 each (cheapest off of newegg) which brings you to a total of $65326.67!! Throw in some money for some chassis and ghetto wiring and I'm sure you can build your own petabyte server for much less!! hahaha
There were two independent large-scale studies of hard-drive failures done a few years ago. Among their common conclusions was that there was no significant difference in reliability between “enterprise-grade” and “consumer-grade” hard drives.
Also, notice they’re using Linux software RAID. That helps to keep the complexity down and the reliability up.
Over 5 years, you'd spend more time and money at that scale moving from pod to pod replacing sata drives that fail and then degraded performance during rebuilds.
The title of this article and the article it uses as it source is clearly a misnomer.
This isn't a cloud storage solution, its a box with a crap ton of storage, that's all. Is it cheap? Yes. Will it function as a cloud storage? Probably as long as you like all your eggs in on basket.
The important part of the puzzle is the replication and storage methodology that actually handles where the data goes. Obviously that is proprietary and I wouldn't expect them to give it away, but this is just a storage node with a bunch of drives.
This system might be the best thing for their particular business. However, it has very poor reliability at the pod level - No redundant anything - and low performance. Is that a bad thing? No. Does that line up to a lot of customers in the external space? No.
Also, I found no mention in the article about how much they spent to design the box. That counts in the overall cost of the system.
They're building this cluster out of cheap, consumer level hardware. Who in their right mind does that with a production server?
Just looking at the hard drive reviews on Newegg people list them as having random problems in RAID configurations.
No thanks. If I'm storing that much data (and presumably it is important data), I'm not going to trust it to some Newegg door buster - free shipping special.
If Google can build the worlds largest clustered computing environment out of garbage, you can build enough redundancy into hardware this cheap to have your cake and not pay as much for it.
You could even mirror these appliances for total redundancy way less than the cost of most large tier storage solutions.
Besides, just because something says "HP" doesn't mean it's not garbage. It often is; it's just well supported garbage.
lol you can build the same if you use 666 1.5 Terabyte drives at $97.99 each (cheapest off of newegg) which brings you to a total of $65326.67!! Throw in some money for some chassis and ghetto wiring and I'm sure you can build your own petabyte server for much less!! hahaha
There were two independent large-scale studies of hard-drive failures done a few years ago. Among their common conclusions was that there was no significant difference in reliability between “enterprise-grade” and “consumer-grade” hard drives.
Also, notice they’re using Linux software RAID. That helps to keep the complexity down and the reliability up.
Over 5 years, you'd spend more time and money at that scale moving from pod to pod replacing sata drives that fail and then degraded performance during rebuilds.
The title of this article and the article it uses as it source is clearly a misnomer.
This isn't a cloud storage solution, its a box with a crap ton of storage, that's all. Is it cheap? Yes. Will it function as a cloud storage? Probably as long as you like all your eggs in on basket.
The important part of the puzzle is the replication and storage methodology that actually handles where the data goes. Obviously that is proprietary and I wouldn't expect them to give it away, but this is just a storage node with a bunch of drives.
I'll take buzz word bingo for $500 Alex.
This system might be the best thing for their particular business. However, it has very poor reliability at the pod level - No redundant anything - and low performance. Is that a bad thing? No. Does that line up to a lot of customers in the external space? No.
Also, I found no mention in the article about how much they spent to design the box. That counts in the overall cost of the system.
They're building this cluster out of cheap, consumer level hardware. Who in their right mind does that with a production server?
Just looking at the hard drive reviews on Newegg people list them as having random problems in RAID configurations.
No thanks. If I'm storing that much data (and presumably it is important data), I'm not going to trust it to some Newegg door buster - free shipping special.