Main Area and Open Discussion > Living Room
The end of the hard disk
Renegade:
Raid 1 mirrored SSDs. :)
They do have good MTBFs.
Still, SAS may still be the most prudent way to go for the moment.
Shades:
Around 1000USD a pop you have now decently sized hard disks that work according to SSD principles, but instead of using SATA, they use the PCI-Express lanes of your motherboard. If you think SSD's (or SSD's in RAID) are fast...these puppies run 4 to 5 times faster than SSD drives (at their top speed) in most usage scenarios. If you want really fast servers that have no problems shifting mountains of data around, SSD's are already old hat.
Renegade:
Around 1000USD a pop you have now decently sized hard disks that work according to SSD principles, but instead of using SATA, they use the PCI-Express lanes of your motherboard. If you think SSD's (or SSD's in RAID) are fast...these puppies run 4 to 5 times faster than SSD drives (at their top speed) in most usage scenarios. If you want really fast servers that have no problems shifting mountains of data around, SSD's are already old hat.
-Shades (June 23, 2015, 11:09 PM)
--- End quote ---
Heh! After both my desktop and laptop died recently, I was looking around and saw those. They sound pretty sweet.
highend01:
Around 1000USD a pop you have now decently sized hard disks that work according to SSD principles, but instead of using SATA, they use the PCI-Express lanes of your motherboard. If you think SSD's (or SSD's in RAID) are fast...these puppies run 4 to 5 times faster than SSD drives (at their top speed) in most usage scenarios. If you want really fast servers that have no problems shifting mountains of data around, SSD's are already old hat.
--- End quote ---
Link please :)
wraith808:
Simple search brings up several links....
http://lmgtfy.com/?q=SSD+use+the+PCI-Express+lanes
One example from that search...
http://www.extremetech.com/computing/202603-intels-new-pcie-ssd-750-brings-enterprise-features-to-the-client-market
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version