Personal computing discussed

Moderators: morphine, Steel

  • 1
  • 3
  • 4
  • 5
  • 6
  • 7
 
Glorious
Gold subscriber
Gerbilus Supremus
Posts: 11661
Joined: Tue Aug 27, 2002 6:35 pm

Re: Blackblaze's 2018 HDD stats

Mon Apr 15, 2019 3:47 pm

Yeah these were things with like 8 or 16 GiB of flash or whatever and like straight from the OEM they had differing block sizes.

It was like that utility could never work.
 
Igor_Kavinski
Gerbil XP
Posts: 477
Joined: Fri Dec 22, 2006 2:34 am

Re: Blackblaze's 2018 HDD stats

Tue Apr 16, 2019 4:28 am

just brew it! wrote:
Waco wrote:
SSDs and thumb drives are still plagued with this, though. I even have 3 thumb drives from the same product line bought in a quantity of 3, all at once, that are 3 different sizes.

...and that's why I qualified my observation with "...on mechanical HDDs..." :wink:

And yeah, I've seen the weird thumb drive capacity discrepancy thing too, on drives of the same model.


That could be explained by flash cells in the chips not quite meeting quality standards and then being disabled. Rather than throw away the whole chip, they probably figure the size difference is acceptable as long as it is "close" to the advertised capacity, give or take several megabytes.
 
just brew it!
Gold subscriber
Administrator
Posts: 52805
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Blackblaze's 2018 HDD stats

Tue Apr 16, 2019 6:07 am

Igor_Kavinski wrote:
That could be explained by flash cells in the chips not quite meeting quality standards and then being disabled. Rather than throw away the whole chip, they probably figure the size difference is acceptable as long as it is "close" to the advertised capacity, give or take several megabytes.

That's kind of scary because it implies there are no/few spare cells. Otherwise they could just map the bad cells to spares up front and make the user-visible size the same.
Nostalgia isn't what it used to be.
 
Usacomp2k3
Gerbil God
Posts: 22843
Joined: Thu Apr 01, 2004 4:53 pm
Location: Orlando, FL
Contact:

Re: Blackblaze's 2018 HDD stats

Tue Apr 16, 2019 6:34 am

just brew it! wrote:
Igor_Kavinski wrote:
That could be explained by flash cells in the chips not quite meeting quality standards and then being disabled. Rather than throw away the whole chip, they probably figure the size difference is acceptable as long as it is "close" to the advertised capacity, give or take several megabytes.

That's kind of scary because it implies there are no/few spare cells. Otherwise they could just map the bad cells to spares up front and make the user-visible size the same.

What value do unused spare cells have? I'd rather the memory size decrease over time as bad cells present themselves than to have memory physically there that I can't take advantage of.
 
just brew it!
Gold subscriber
Administrator
Posts: 52805
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Blackblaze's 2018 HDD stats

Tue Apr 16, 2019 7:20 am

Usacomp2k3 wrote:
just brew it! wrote:
Igor_Kavinski wrote:
That could be explained by flash cells in the chips not quite meeting quality standards and then being disabled. Rather than throw away the whole chip, they probably figure the size difference is acceptable as long as it is "close" to the advertised capacity, give or take several megabytes.

That's kind of scary because it implies there are no/few spare cells. Otherwise they could just map the bad cells to spares up front and make the user-visible size the same.

What value do unused spare cells have? I'd rather the memory size decrease over time as bad cells present themselves than to have memory physically there that I can't take advantage of.

The problem with this is that if you format the file system based on the larger size, the OS will not automatically shrink the filesystem if the underlying device shrinks. AFAIK none of the widely used file systems (WIndows or *NIX) dynamically resize file systems on the fly if the underlying device changes size, because block devices like HDDs, SSDs, and USB thumbdrives simply do not do that. You'll just have a corrupted file system since the meta-data which describes the file system won't match the size of the underlying device.

Furthermore, the way "bad cells present themselves" will be that you'll have a file or folder that you simply can't read back (i.e. data loss). Unless you want the OS to test the entire device for bad blocks at format time, which (given the slow, bottom-of-the-barrel flash used in many thumb drives) could take hours.

What you're suggesting isn't practical for the reasons noted above.
Nostalgia isn't what it used to be.
 
Usacomp2k3
Gerbil God
Posts: 22843
Joined: Thu Apr 01, 2004 4:53 pm
Location: Orlando, FL
Contact:

Re: Blackblaze's 2018 HDD stats

Tue Apr 16, 2019 8:52 am

just brew it! wrote:
Usacomp2k3 wrote:
What value do unused spare cells have? I'd rather the memory size decrease over time as bad cells present themselves than to have memory physically there that I can't take advantage of.

The problem with this is that if you format the file system based on the larger size, the OS will not automatically shrink the filesystem if the underlying device shrinks. AFAIK none of the widely used file systems (WIndows or *NIX) dynamically resize file systems on the fly if the underlying device changes size, because block devices like HDDs, SSDs, and USB thumbdrives simply do not do that. You'll just have a corrupted file system since the meta-data which describes the file system won't match the size of the underlying device.

Good point. I hadn't thought of that.
just brew it! wrote:
Furthermore, the way "bad cells present themselves" will be that you'll have a file or folder that you simply can't read back (i.e. data loss). Unless you want the OS to test the entire device for bad blocks at format time, which (given the slow, bottom-of-the-barrel flash used in many thumb drives) could take hours.

What you're suggesting isn't practical for the reasons noted above.

How does it do it when there is provisioning? I would think the process would be the same (other than the issue you mentioned above about dynamically resizing the filesystem.
 
Waco
Gold subscriber
Minister of Gerbil Affairs
Posts: 2936
Joined: Tue Jan 20, 2009 4:14 pm
Location: Los Alamos, NM

Re: Blackblaze's 2018 HDD stats

Tue Apr 16, 2019 11:20 am

Usacomp2k3 wrote:
How does it do it when there is provisioning? I would think the process would be the same (other than the issue you mentioned above about dynamically resizing the filesystem.

If things are working properly, the controller on the SSD will remap bad sectors as they wear out before data loss occurs. The device size stays constant.

HDDs do the same thing.
Desktop: Z170A | 6700K @ 4.4 | 32 GB | Alphacool Eisblock Radeon VII | Heatkiller R3 | Samsung 4K 40" | 1 TB NVME + 2 TB SATA + LSI (128x8) RAID 0
NAS: 1950X | Designare EX | 32 GB ECC | 7x8 TB RAIDZ2 | 8x2 TB RAID10 | FreeNAS | ZFS | LSI SAS
 
just brew it!
Gold subscriber
Administrator
Posts: 52805
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Blackblaze's 2018 HDD stats

Tue Apr 16, 2019 10:33 pm

Waco wrote:
Usacomp2k3 wrote:
How does it do it when there is provisioning? I would think the process would be the same (other than the issue you mentioned above about dynamically resizing the filesystem.

If things are working properly, the controller on the SSD will remap bad sectors as they wear out before data loss occurs. The device size stays constant.

HDDs do the same thing.

HDDs sort of do that in theory, but as often as not the remapping doesn't occur until you've actually lost a block of data. Permanent read error causes the block to go to "pending reallocation" status; next write to that block will cause a write-read-verify, and if the verify fails it gets redirected to a spare sector.
Nostalgia isn't what it used to be.
 
Igor_Kavinski
Gerbil XP
Posts: 477
Joined: Fri Dec 22, 2006 2:34 am

Re: Blackblaze's 2018 HDD stats

Wed Apr 17, 2019 6:42 am

Something new for data hoarders to worry about: bitflux!

Learned about this here: https://blog.macsales.com/43702-we-bet- ... ercise-too

Refreshing instructions for Mac users: https://larryjordan.com/articles/techni ... k-storage/

http://www.puransoftware.com/DiskFresh.html

https://superuser.com/questions/514035/ ... ith-chkdsk

So a question: what do data center admins do for bitflux prevention?
 
just brew it!
Gold subscriber
Administrator
Posts: 52805
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Blackblaze's 2018 HDD stats

Wed Apr 17, 2019 7:19 am

SSDs which are powered on should refresh the data if the charge levels in the flash cells are getting low, much as DRAM cells get refreshed (just on a timescale of weeks or months instead of milliseconds). Not to say they all do this, but the better ones should. SSDs which are powered off could have issues if they are near the end of their endurance rating, which will shorten data retention times.

Refreshing a SSD more than necessary is actually harmful, since you will eat into the write endurance of the flash cells.

HDDs which are powered down have historically been very good at retaining data over multiple years, though this may be getting shorter with today's high bit and track densities, since this closer packing of bits increases the effect that adjacent bits' magnetic fields have on each other over time, and also blurs the boundaries between the individual bits making them harder to read back even in ideal cases. There's no benefit to leaving a HDD powered on since AFAIK none of them automatically refresh data.

Datacenters often employ a two-pronged approach to dealing with failing media and "bitrot": Redundancy (mirroring, parity, or erasure coding) and scrubbing (periodically sweeping through all of the data, verifying its integrity using whatever redundancy mechanism is being employed), and rewriting any missing/corrupted data with a fresh copy. Home and small business users can employ the same ideas on a smaller scale, since the better NAS boxes (and Linux software RAID) allow you to do things like RAID-5/6 with scrubbing.
Nostalgia isn't what it used to be.
 
Waco
Gold subscriber
Minister of Gerbil Affairs
Posts: 2936
Joined: Tue Jan 20, 2009 4:14 pm
Location: Los Alamos, NM

Re: Blackblaze's 2018 HDD stats

Wed Apr 17, 2019 11:45 am

just brew it! wrote:
HDDs sort of do that in theory, but as often as not the remapping doesn't occur until you've actually lost a block of data. Permanent read error causes the block to go to "pending reallocation" status; next write to that block will cause a write-read-verify, and if the verify fails it gets redirected to a spare sector.

I know the specs don't say it, but I have pretty good reason to believe that certain drives will remap sectors on the fly *before* the internal ECC can't correct for it. They don't increment the reallocated sectors when that happens.
Desktop: Z170A | 6700K @ 4.4 | 32 GB | Alphacool Eisblock Radeon VII | Heatkiller R3 | Samsung 4K 40" | 1 TB NVME + 2 TB SATA + LSI (128x8) RAID 0
NAS: 1950X | Designare EX | 32 GB ECC | 7x8 TB RAIDZ2 | 8x2 TB RAID10 | FreeNAS | ZFS | LSI SAS
 
meerkt
Graphmaster Gerbil
Topic Author
Posts: 1174
Joined: Sun Aug 25, 2013 2:55 am

Re: Blackblaze's 2018 HDD stats

Thu Apr 18, 2019 12:34 pm

just brew it! wrote:
That's kind of scary because it implies there are no/few spare cells. Otherwise they could just map the bad cells to spares up front and make the user-visible size the same.

I'm guessing they want to have n% spare when the user gets it, so they eat into the user-addressable space.
And indeed it's very annoying, and perhaps suggestive of poor quality flash. I try to avoid models or manufacturers that do that, if I can.

just brew it! wrote:
This would cause problems in RAID configurations if you tried to add a drive to an array that was originally built with slightly larger drives.

The simple solution is to leave out 10-100MB when doing the initialization.

just brew it! wrote:
The problem with this is that if you format the file system based on the larger size, the OS will not automatically shrink the filesystem if the underlying device shrinks.

No need to resize the partition, just flag bad clusters or sectors. Even DOS could do that. :)

Waco wrote:
reason to believe that certain drives will remap sectors on the fly *before* the internal ECC can't correct for it. They don't increment the reallocated sectors when that happens.

Which models? Remapping early is good. But are you sure they don't increment the reallocated count? That beats the whole idea of SMART.
 
just brew it!
Gold subscriber
Administrator
Posts: 52805
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Blackblaze's 2018 HDD stats

Thu Apr 18, 2019 12:43 pm

meerkt wrote:
just brew it! wrote:
The problem with this is that if you format the file system based on the larger size, the OS will not automatically shrink the filesystem if the underlying device shrinks.

No need to resize the partition, just flag bad clusters or sectors. Even DOS could do that. :)

The problem with this is twofold:
1. If you want to test for bad sectors at format time, then formatting takes too long.
2. Marking out bad sectors on the fly at the file system level when a read error occurs implies you've already lost data.
Nostalgia isn't what it used to be.
 
meerkt
Graphmaster Gerbil
Topic Author
Posts: 1174
Joined: Sun Aug 25, 2013 2:55 am

Re: Blackblaze's 2018 HDD stats

Thu Apr 18, 2019 1:09 pm

No point in exposing bad sectors the manufacturer already detected, only those found on the fly.

For those, I do wonder if NVMe/AHCI/ATA can give more result codes than just good and bad.
A read from a borderline page could return the data + flag to tell the OS it should mark it as bad.

It might be doable also without it, as long as there are some free pages (TRIM-based). The drive could remap internally, and on the next write to a remap-target page, return "bad".
 
Waco
Gold subscriber
Minister of Gerbil Affairs
Posts: 2936
Joined: Tue Jan 20, 2009 4:14 pm
Location: Los Alamos, NM

Re: Blackblaze's 2018 HDD stats

Thu Apr 18, 2019 3:25 pm

meerkt wrote:
Which models? Remapping early is good. But are you sure they don't increment the reallocated count? That beats the whole idea of SMART.

I can't say specifically, but let's go with "large drives" with 4KN formatting.
Desktop: Z170A | 6700K @ 4.4 | 32 GB | Alphacool Eisblock Radeon VII | Heatkiller R3 | Samsung 4K 40" | 1 TB NVME + 2 TB SATA + LSI (128x8) RAID 0
NAS: 1950X | Designare EX | 32 GB ECC | 7x8 TB RAIDZ2 | 8x2 TB RAID10 | FreeNAS | ZFS | LSI SAS
 
just brew it!
Gold subscriber
Administrator
Posts: 52805
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Blackblaze's 2018 HDD stats

Thu Apr 18, 2019 9:20 pm

meerkt wrote:
No point in exposing bad sectors the manufacturer already detected, only those found on the fly.

For those, I do wonder if NVMe/AHCI/ATA can give more result codes than just good and bad.
A read from a borderline page could return the data + flag to tell the OS it should mark it as bad.

It might be doable also without it, as long as there are some free pages (TRIM-based). The drive could remap internally, and on the next write to a remap-target page, return "bad".

If the drive is part of a RAID array, returning a failure on write is probably gonna cause the RAID controller/software to kick it from the array.
Nostalgia isn't what it used to be.
 
Waco
Gold subscriber
Minister of Gerbil Affairs
Posts: 2936
Joined: Tue Jan 20, 2009 4:14 pm
Location: Los Alamos, NM

Re: Blackblaze's 2018 HDD stats

Thu Apr 18, 2019 10:22 pm

just brew it! wrote:
If the drive is part of a RAID array, returning a failure on write is probably gonna cause the RAID controller/software to kick it from the array.

Yet another reason why ZFS is superior in many ways to traditional RAID arrays. :)
Desktop: Z170A | 6700K @ 4.4 | 32 GB | Alphacool Eisblock Radeon VII | Heatkiller R3 | Samsung 4K 40" | 1 TB NVME + 2 TB SATA + LSI (128x8) RAID 0
NAS: 1950X | Designare EX | 32 GB ECC | 7x8 TB RAIDZ2 | 8x2 TB RAID10 | FreeNAS | ZFS | LSI SAS
 
Topinio
Gerbil Jedi
Posts: 1733
Joined: Mon Jan 12, 2015 9:28 am
Location: London

Re: Blackblaze's 2018 HDD stats

Fri Apr 19, 2019 5:34 pm

Waco wrote:
I can't say specifically, but let's go with "large drives" with 4KN formatting.

Oh, come on.

You're only alluding to enterprise high-end drives anyway, at least be prepared to either hold your tongue or spill the beans, this "I know but won't say" stuff is lame.
Desktop: E3-1270 v5, X11SAT-F, 32GB, RX Vega 56, 500GB Crucial P1, 250GB Crucial MX500, 4TB 7E8, Xonar DGX, XL2730Z + G2420HDB
HTPC: i5-2500K, DH67GD, 6GB, RX 580, 250GB MX500, 1.5TB Barracuda, Xonar DX
Laptop: MacBook6,1
 
Waco
Gold subscriber
Minister of Gerbil Affairs
Posts: 2936
Joined: Tue Jan 20, 2009 4:14 pm
Location: Los Alamos, NM

Re: Blackblaze's 2018 HDD stats

Fri Apr 19, 2019 9:46 pm

Topinio wrote:
Oh, come on.

You're only alluding to enterprise high-end drives anyway, at least be prepared to either hold your tongue or spill the beans, this "I know but won't say" stuff is lame.

I'm sorry, would you rather I say nothing at all? You can figure out pretty easily what I'm talking about and I'd rather not be overly specific.

And no, I'm not talking about enterprise drives specifically.
Desktop: Z170A | 6700K @ 4.4 | 32 GB | Alphacool Eisblock Radeon VII | Heatkiller R3 | Samsung 4K 40" | 1 TB NVME + 2 TB SATA + LSI (128x8) RAID 0
NAS: 1950X | Designare EX | 32 GB ECC | 7x8 TB RAIDZ2 | 8x2 TB RAID10 | FreeNAS | ZFS | LSI SAS
 
just brew it!
Gold subscriber
Administrator
Posts: 52805
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Blackblaze's 2018 HDD stats

Fri Apr 19, 2019 11:02 pm

@Topinio - Ever heard of NDAs? Some of us work in the industry and need to dance around that issue. While I was not aware of what Waco is alluding to, there's other HDD industry related inside info (a.k.a. "dirty laundry") I've encountered through the course of my day job that I would've loved to post; but my desire to keep my job takes precedence.
Nostalgia isn't what it used to be.
 
Waco
Gold subscriber
Minister of Gerbil Affairs
Posts: 2936
Joined: Tue Jan 20, 2009 4:14 pm
Location: Los Alamos, NM

Re: Blackblaze's 2018 HDD stats

Fri Apr 19, 2019 11:40 pm

just brew it! wrote:
...but my desire to keep my job takes precedence.

Ditto. I can not post anything if many would prefer that, but I figured a tidbit is better than nothing especially if reading between the lines is simple.
Desktop: Z170A | 6700K @ 4.4 | 32 GB | Alphacool Eisblock Radeon VII | Heatkiller R3 | Samsung 4K 40" | 1 TB NVME + 2 TB SATA + LSI (128x8) RAID 0
NAS: 1950X | Designare EX | 32 GB ECC | 7x8 TB RAIDZ2 | 8x2 TB RAID10 | FreeNAS | ZFS | LSI SAS
 
Topinio
Gerbil Jedi
Posts: 1733
Joined: Mon Jan 12, 2015 9:28 am
Location: London

Re: Blackblaze's 2018 HDD stats

Sat Apr 20, 2019 6:18 am

Waco wrote:
I'm sorry, would you rather I say nothing at all? You can figure out pretty easily what I'm talking about and I'd rather not be overly specific.

Sorry, dunno why that rubbed me up the wrong way yesterday.

And I may be being thick, but I didn't know what drives are 4kN and remap sectors before the ECC can't correct for it without incrementing the reallocated sectors count...

just brew it! wrote:
@Topinio - Ever heard of NDAs?

Yeah, ofc, but ... if what I want to say would break NDA, I wouldn't say it.

Ignore me!
Desktop: E3-1270 v5, X11SAT-F, 32GB, RX Vega 56, 500GB Crucial P1, 250GB Crucial MX500, 4TB 7E8, Xonar DGX, XL2730Z + G2420HDB
HTPC: i5-2500K, DH67GD, 6GB, RX 580, 250GB MX500, 1.5TB Barracuda, Xonar DX
Laptop: MacBook6,1
 
just brew it!
Gold subscriber
Administrator
Posts: 52805
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Blackblaze's 2018 HDD stats

Sat Apr 20, 2019 6:56 am

Topinio wrote:
just brew it! wrote:
@Topinio - Ever heard of NDAs?

Yeah, ofc, but ... if what I want to say would break NDA, I wouldn't say it.

Right. So the choice is either to say nothing at all, or to speak in generalities without naming manufacturers or models.
Nostalgia isn't what it used to be.
 
meerkt
Graphmaster Gerbil
Topic Author
Posts: 1174
Joined: Sun Aug 25, 2013 2:55 am

Re: Blackblaze's 2018 HDD stats

Sat Apr 20, 2019 7:22 am

just brew it! wrote:
If the drive is part of a RAID array, returning a failure on write is probably gonna cause the RAID controller/software to kick it from the array.

I was thinking more of home use, and SSDs. I guess that's what Usacomp2k3 meant. (I just noticed the forum's oddly missing a jump-to-quoted-message feature.)

And again the old idea: it'd be nice if drives allowed configurable operation modes instead of separate models for consumer/NAS/whatever.
Old ATA already supports a configurable hidden area (HPA). It could be extended to allow spare area tweaks. Who knows, maybe it's already possible.
 
just brew it!
Gold subscriber
Administrator
Posts: 52805
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Blackblaze's 2018 HDD stats

Sat Apr 20, 2019 7:28 am

meerkt wrote:
And again the old idea: it'd be nice if drives allowed configurable operation modes instead of separate models for consumer/NAS/whatever.

FWIW some of the Toshiba drives I own allow TLER to be toggled on/off by the user.
Nostalgia isn't what it used to be.
 
Waco
Gold subscriber
Minister of Gerbil Affairs
Posts: 2936
Joined: Tue Jan 20, 2009 4:14 pm
Location: Los Alamos, NM

Re: Blackblaze's 2018 HDD stats

Mon Apr 22, 2019 12:18 am

just brew it! wrote:
Right. So the choice is either to say nothing at all, or to speak in generalities without naming manufacturers or models.

Exactly. I would like to think that *some* idea of what is going on is useful as opposed to nothing.

It's a little annoying from the outside, but as someone who gobbled up things like this in the early/mid aughts, I know I learned a lot and appreciated the information.

just brew it! wrote:
meerkt wrote:
And again the old idea: it'd be nice if drives allowed configurable operation modes instead of separate models for consumer/NAS/whatever.

FWIW some of the Toshiba drives I own allow TLER to be toggled on/off by the user.

Hm, I didn't know that. Given I run ZFS on damn near everything I don't have much stake in it but for those that run "real" RAIDs these days that's super useful.
Desktop: Z170A | 6700K @ 4.4 | 32 GB | Alphacool Eisblock Radeon VII | Heatkiller R3 | Samsung 4K 40" | 1 TB NVME + 2 TB SATA + LSI (128x8) RAID 0
NAS: 1950X | Designare EX | 32 GB ECC | 7x8 TB RAIDZ2 | 8x2 TB RAID10 | FreeNAS | ZFS | LSI SAS
 
JustAnEngineer
Gold subscriber
Gerbil God
Posts: 18612
Joined: Sat Jan 26, 2002 7:00 pm
Location: The Heart of Dixie

Re: Blackblaze's 2018 HDD stats

Mon May 06, 2019 4:23 am

https://www.techpowerup.com/255181/hard ... oy-in-2019
Japanese company Nidec, which manufactures nearly 85% of all DC motors for use in HDDs across the industry, estimates a nearly 50 percent drop in HDD shipments for 2019.
i7-9700K, NH-D15, Z390M Pro4, 32 GiB, RX Vega64, Define Mini-C, SSR-850PX, C32HG70+U2410, RK-9000BR, MX518
 
just brew it!
Gold subscriber
Administrator
Posts: 52805
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: Blackblaze's 2018 HDD stats

Mon May 06, 2019 5:35 am

JustAnEngineer wrote:
https://www.techpowerup.com/255181/hard-drive-shipments-expected-to-drop-nearly-50-percent-yoy-in-2019
Japanese company Nidec, which manufactures nearly 85% of all DC motors for use in HDDs across the industry, estimates a nearly 50 percent drop in HDD shipments for 2019.

Not a huge surprise. And nearly all of that remaining volume will be going to Cloud storage providers and large enterprise customers.

It would be interesting to see what the relative capacity numbers (in total TB shipped) are YoY. Since Cloud/enterprise users are going to favor the highest capacity models, the total storage capacity shipped probably won't drop much, or could even rise.
Nostalgia isn't what it used to be.
 
Waco
Gold subscriber
Minister of Gerbil Affairs
Posts: 2936
Joined: Tue Jan 20, 2009 4:14 pm
Location: Los Alamos, NM

Re: Blackblaze's 2018 HDD stats

Mon May 06, 2019 9:45 am

It'll likely rise again, as usual. Manufacturers have been dropping the older/smaller models from production faster than any other time in history. I can't even get commitments from vendors that drives that just started shipping will be available in a couple years. The enterprise market used to keep the older/smaller models alive for half a decade or more, but there's an aggressive push to only make the biggest drives possible at any point in time.

Cloud is certainly pushing things in interesting directions.
Desktop: Z170A | 6700K @ 4.4 | 32 GB | Alphacool Eisblock Radeon VII | Heatkiller R3 | Samsung 4K 40" | 1 TB NVME + 2 TB SATA + LSI (128x8) RAID 0
NAS: 1950X | Designare EX | 32 GB ECC | 7x8 TB RAIDZ2 | 8x2 TB RAID10 | FreeNAS | ZFS | LSI SAS
 
Krogoth
Gold subscriber
Gerbil Elder
Posts: 5716
Joined: Tue Apr 15, 2003 3:20 pm
Location: somewhere on Core Prime
Contact:

Re: Blackblaze's 2018 HDD stats

Mon May 06, 2019 10:50 am

Not a shocker at all.

Flash media is starting to become cheap enough for OEMs to considered a the default option in bargain basement desktop/laptop SKUs. M.2 Express ports are ubiquitous on modern platforms and make it easier to engineer SFF PCs/ultra-portable laptops.
Gigabyte Z390 AORUS-PRO Coffee Lake R 9700K, 2x8GiB of G.Skill DDR4-3600, Sapphire RX Vega 64, Corsair CX-750M V2 and Fractal Define R4 (W)
Ivy Bridge 3570K, 2x4GiB of G.Skill RIPSAW DDR3-1600, Gigabyte Z77X-UD3H, Corsair CX-750M V2, and PC-7B
  • 1
  • 3
  • 4
  • 5
  • 6
  • 7

Who is online

Users browsing this forum: No registered users and 15 guests