Seagate Launches the Final 15K rpm Hard Drive: RIP 15K HDDs

[Patrick Kennedy]

The Seagate Enterprise Performance 15K.6 is the last 15k rpm hard drive series we foresee launched as SSD technology is two (almost three) generations ahead

The post Seagate Launches the Final 15K rpm Hard Drive: RIP 15K HDDs appeared first on ServeTheHome.

Related posts:

1. Seagate Nytro XF1230 SATA SSD Released
2. Seagate Nytro XP7102 NVMe Add-in Card Released
3. Toshiba launches three new NVMe SSD series
4. Seagate 10TB drives now shipping in volume

Continue reading...

 

[wildpig1234]

when are we going to ever see the inflection point for ssd when it comes to regular consumer market? My guess is never?

 

[mstone]

when are we going to ever see the inflection point for ssd when it comes to regular consumer market? My guess is never?

When the volume on the spinners drops low enough that they don't enjoy economies of scale. Some day you'll wake up and they'll just be gone. I think we're already past the point where spinning disks would be cost competitive if it weren't for pre-existing investments in the equipment to produce them.

 

[wildpig1234]

The different demands of high i/o and other performance bench in the server market is probably what kill the spinners. consumer market doesn't demand the high i/o and fast transfer rate so i think that's why spinners will be around for quite a while in consumer market because it is "good" enough.

 

[mstone]

The different demands of high i/o and other performance bench in the server market is probably what kill the spinners. consumer market doesn't demand the high i/o and fast transfer rate so i think that's why spinners will be around for quite a while in consumer market because it is "good" enough.

The consumer market is driven almost entirely by cost. A thinking consumer will prefer an SSD because: quieter, lower power (longer battery life), more reliable, basic OS operations are noticeably faster. So why would any consumer want a spinner? Cost is the only reason--and consider carefully whether in the long run it costs more to fab and solder up a few chips for an SSD or assemble chips, motor, platters, heads, servos, etc for a spinner. As production volume for SSD storage continues to go up, it's basically inevitable that they'll end up cheaper than spinners.

 

[ttabbal]

For a lot of needs though, consumers need cost/TB to match up. For things like backups, archival stuff like photos and movies, they just need too much space for SSD to be practical. And those uses don't need much speed.

I agree that the sunk cost of manufacturing spinners is the only reason they can compete at all though. If they had to tool up for spinners or SSD, there's no contest, SSD would be cheaper. But they already own all that, so it's mostly incremental upgrades here and there to stay relevant. Little increases in density here and there get it done. Might as well keep using up that inventory of motors, etc..

I still own more spinners than SSD. In a big enough array, spinners are fast enough to be useful. And the cost/TB is excellent. $30/2TB simply can't be done in SSD right now. Eventually, but not yet.

I'm a little surprised they are making a new 15K spinner. I can't imagine new enterprise gear using them when enterprise SSD is probably cost competitive with this. I can only figure they have a few large customers that want them, and they already have inventory of 15K motors to use up.

 

[cheezehead]

Consumer space is very different in terms of workloads vs enterprise space. Cost is not the only factor in either space.

For consumers cost/TB a big factor but not also not really. when people buy new systems, they are looking at the total system cost....not very often the drive cost alone. Consumer SSD's can almost be considered a WORM workload. In which case n-layer TLC-based SSDs with a WPD rating of under 0.3 still work very well. In this scenario, the cost difference is not much today and will disappear sooner than later as more and more laptops are doing m.2 slots only.

For enterprise 15k drives hopefully will die in the next few years, I'm assuming the main desire for for enterprise customers looking to added "performance" drives to existing arrays. This is something that over the next few years will go by the wayside as customers will purchasing either all flash arrays or tiered arrays mixing different classes of SSDs with high capacity spinners for an archival tier.