On Fri, 3 Apr 2015 11:16:11 +0300 Andrey Korolyov wrote: > On Thu, Apr 2, 2015 at 8:03 PM, Mark Nelson <mnelson@xxxxxxxxxx> wrote: > > Thought folks might like to see this: > > > > http://hothardware.com/reviews/intel-ssd-750-series-nvme-pci-express-solid-state-drive-review > > > > Quick summary: > > > > - PCIe SSD based on the P3700 > > - 400GB for $389! > > - 1.2GB/s writes and 2.4GB/s reads > > - power loss protection > > - 219TB write endurance > > > > So basically looks extremely attractive on paper except for the write > > endurance. I suspect this is not actually using HET cells (a summary > > I read said it was). How far beyond Intel's endurance rating the card > > can go is the big question. > > > > Mark > > _______________________________________________ > > ceph-users mailing list > > ceph-users@xxxxxxxxxxxxxx > > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com > > All the characteristics are awesome except resource - it would burn > out in a couple of months and replacement is coupled with node > poweroff... > While I basically agree, I'd point out that these also come in a 2.5" MVNE version, so that's hot-swappable, see some current SuperMicro servers for example. Also it really depends on a number of factors, I've got a cluster that racks up about 40TB/year per SSD journal (2 of them per node), so it would be (barely) within the warranty limitations. But of course people who'd actually want/need that kind of performance are probably writing a lot more than what this cluster gets, thus burning them out faster. Balancing these SSDs with regards to backing storage (and CPU power) seems rather tricky to me. Lastly, the fact that they "die" precisely at the time the TBW is exceed makes them particularly unattractive when compared to the same sized and priced 3610 that is 15 times cheaper in TBW/$. Regards, Christian -- Christian Balzer Network/Systems Engineer chibi@xxxxxxx Global OnLine Japan/Fusion Communications http://www.gol.com/ _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
 |