When do you think we will see Helium filled 2.5" hard drives?

cbn

Lifer
Mar 27, 2009
12,968
221
106
The 3.5" form factor has gotten all Helium development so far, but with MAMR tech coming and NAND approaching its limits on scaling (128 layers before string stacking) I'm thinking the hard drive companies will be more likely to use Helium on 2.5" as well.

So how do you think they will proceed first?

15mm first? Then 9.5mm? Then 7mm? Or a different order?

How many platters?

5400 rpm only? or 7200 rpm also?
 
Last edited:

Elixer

Lifer
May 7, 2002
10,371
762
126
I think laptop builders will drop SATA, and instead have multiple M.2 slots for storage, except for the very low end.
Right now, I think the max limit is 3 platters for a 2.5" devices, and that should still be the max even if they use He, since they still need room for the seal & laser weld them shut.
Using He also lowers the power required for the HD to spin them, so, it is possible we will see some other speed other than 5400 or 7200RPM.
 
  • Like
Reactions: cbn

cbn

Lifer
Mar 27, 2009
12,968
221
106
Right now, I think the max limit is 3 platters for a 2.5" devices, and that should still be the max even if they use He, since they still need room for the seal & laser weld them shut.

There actually is a 2.5" 15mm HDD with five platters:

https://www.anandtech.com/show/1075...5-mobile-hard-drives-with-up-to-5-tb-capacity

And here is another 2.5" 15mm HDD with four platters:

https://www.kitguru.net/components/...-drive-with-3tb-capacity-finally-hits-retail/

I was thinking maybe they would go to 7 platters (for 15mm) if using Helium...but I don't how much a difference thinner platters would make in this scenario.
 
Last edited:

whm1974

Diamond Member
Jul 24, 2016
9,436
1,569
126
There actually is a 2.5" 15mm HDD with five platters:

https://www.anandtech.com/show/1075...5-mobile-hard-drives-with-up-to-5-tb-capacity

And here is another 2.5" 15mm HDD with four platters:

https://www.kitguru.net/components/...-drive-with-3tb-capacity-finally-hits-retail/

I was thinking maybe they would go to 7 platters (for 15mm) if using Helium...but I don't how much a difference thinner platters would make in this scenario.
Any modern laptops they will fit into? I thought that 2.5" drives were all 9mm and now 7mm? Or is that just SSDs?
 
  • Like
Reactions: cbn

cbn

Lifer
Mar 27, 2009
12,968
221
106
Any modern laptops they will fit into?

I am not aware of of any. Even these rather thick laptops use 9.5mm.

15mm is for usb hard drives (ie, external drives) or for 2U Storage (which holds 24 x 2.5" 15mm or 12 x 3.5" drives or 48 2.5" 7mm drives)

I thought that 2.5" drives were all 9mm and now 7mm? Or is that just SSDs?

SSDs is 2.5" x 7mm, but there are some hard drives with 9.5mm form factor.

P.S. PS4/PS4 Pro use 9.5mm.
 
Last edited:

cbn

Lifer
Mar 27, 2009
12,968
221
106
Regarding the 2U Storage, I also found out there are racks that hold 32 x 2.5" x 9.5mm.

So 2U Storage=

48 x 2.5" x 7mm
32 x 2.5" x 9.5mm
24 x 2.5" x 15mm
12 x 3.5"

P.S. I wonder if there is also a 2U rack that holds the thinner 3.5" drives?
 

cbn

Lifer
Mar 27, 2009
12,968
221
106
Back Blaze POD 6.0 holds 60 3.5" drives (top loading) in 4U:

https://www.backblaze.com/b2/storage-pod.html

(This is an impressive 2.5x increase compared to the 24 3.5" drives possible in 2 x 2U 3.5" front loading servers. See previous post)

blog-60-drives-ooh-aah.jpg


So I got to thinking why is there not a 2.5" top loader? Since each 2.5" drive is only 100.35mm long (compared to 146.99mm long for 3.5") the top loader needs to be only 3U right? (Compared to 4U for 3.5" top loader).

Furthermore, since each 2.5" HDD is also narrower (69.85mm vs. 101.6mm) and thinner (15mm vs. 26.11mm) compared to 3.5" I reckon the density increase for 2.5" 15mm top loader (over 3.5" top loader) would be a lot greater than the 2:1 ratio we see with front loading 2U storage.
 
Last edited:

cbn

Lifer
Mar 27, 2009
12,968
221
106
Besides Dense storage, another thing I am wondering about is Corporate VMs.

2.5" 7mm Helium SSHD vs. 2.5" SATA or NVMe SSD?

With enough NAND (say 64 GB on each SSHD) could 2.5" 7mm Helium work for this? This perhaps with two or three platters going as fast as they can (reasonably) with 5v power.

NOTE: I am thinking about 7mm here because 2 x 7mm drives would have twice the IOPS per TB compared to one 15mm drive. This assuming the spindle speed and platter density is the same and the 15mm drive has 2 x the platters of the 7mm.

P.S. Another possibilty for SSHD would be 3DXpoint. but I don't think if we will ever see this with SATA. Maybe eventually with SATA Express? Or NVMe?
 
Last edited:

cbn

Lifer
Mar 27, 2009
12,968
221
106
The story on how Back Blaze built the POD 6.0:

https://www.backblaze.com/blog/open-source-data-storage-server/ ( As it is an open source design this article includes 2-D Blueprints, 3-D Solidwork files, STEP files, Drive Guide design files, Storage Pod 6.0 Parts list, Wiring Diagrams, Wiring routes, and a really good build book which has some very detailed pictures in it)

http://www.storagereview.com/backblaze_storage_pod_60_review

Backblaze Storage Pod 6.0 Specifications

  • Processor: Intel Xeon E5-1620 v2 Quad-Core 3.7GHz 0GT/s 10MB LGA 2011 CPU
  • RAM: HYNIX 4x8GB DDR3-133MHz HMT31GR7CFR4C-PB
  • Motherboard: SuperMicro MBD-X9SRH-7TF-O
  • Power Supply: 2 EVGA 750w
  • Boot Drive: Samsung ST500LM012
  • SATA III Card: Sunrich 4 Port PCI Express A-540
  • Backplane: Sunrich S-331
  • CPU Fan: Dynatron R13 1U
  • Operating System: Linux Debian 7

Data to the 60 3.5" comes from three Sunrich A-540 4 Port PCIe 2.0 x 2 cards. The twelve total SATA 6 Gbps ports (from the three SATA cards) are then multiplied by twelve Sunrich S-331 Backplanes (each SATA port becomes five SATA ports).

The Backblaze Storage Pod 6.0 uses 12 SATA multipliers, each with five ports, for connectivity with the storage drives. Backplane SATA connectivity is provided via 3 SATA III cards with Marvell 9235 chipsets, and the multipliers use Marvell 9715 chipsets.

sunrich-a-540-sata-card_large.png


sunrich-s-331-backplane_large.png


backplane-tray-2.jpg


drive-guide-slides.jpg



So I got to wondering what would be needed for a 2.5" drive version of this?

Assuming 3x as many drives work in a 2.5" x 15mm HDD 3U version of this 3.5" HDD 4U Server that would be 180 drives. (This comes out to be 2520 2.5" drives in 42U vs. 600 3.5" drives in 40U. So exactly 4 x 2.5" drives for every 1 x 3.5" drive via extrapolating 630 3.5" for 42U. With each 2.5" 15mm 5400 rpm drive using about 1/4 power as a 3.5" 7200 rpm drive the power density of 2.5" top loader works out to be the same as 3.5" top loader)

With that noted, what would be the best way to accomplish a 2.5" conversion?

Besides redesigning the drive mounting and changing depth from 4U to 3U, there is the issue of the SATA cards and port multipliers.

Keep the same SATA card arrangement and use port multipliers that turn one SATA into fifteen?

Or should more SATA cards and/or SATA cards with more ports be used and keep port multiplication at 1:5? (Keep in mind no matter approach is taken new backplanes (with the SATA port multipliers) will be needed due to the 2.5" having different drive spacing)
 
Last edited:
  • Like
Reactions: dlerious

cbn

Lifer
Mar 27, 2009
12,968
221
106
Here are the specs on the EVGA 750W PSU that POD 6.0 uses:

https://www.evga.com/products/Specs/PSU.aspx?pn=dbeb37b5-f39d-4754-8378-46ec2f983435

Notice only 25 amps on the 5v rail. So two of them only have 50 amps combined.

This is not going to be enough to run 180 2.5" x 15mm drives (assuming the HDDs cannot use 12v).

P.S. Looking back on POD 3.0, I noticed Back Blaze mentions even the 3.5" Hard drives use a lot of 5v power:

https://www.backblaze.com/blog/180tb-of-good-vibrations-storage-pod-3-0/

In particular, most high end power supplies are designed to deliver most of their power on the 12V rail because that is what high end gamer PCs use. Unfortunately, hard drives draw a lot of power off the 5V rail and can easily overwhelm a high wattage power supply.

Any ideas?

EDIT: Maybe something like this could be added to (built into) the Backplanes?
 
Last edited:
  • Like
Reactions: dlerious

thecoolnessrune

Diamond Member
Jun 8, 2005
9,673
583
126
The average capacity drive has been decreasing the amount of 5V power required, but a high end estimate is about 0.6 Amps of 5V power per drive. At 60 drives, you're at 36 Amps of 5V power, or about 200 watts of 5V power (with margin). A high performance drive might use 0.5 Amps of 5V power in 2.5" form factor. The reason they do the above is because commodity power supplies like that are dirt cheap. If you needed a dedicated PSU, there are lots of manufacturers like Meanwell that produce dedicated voltage PSUs that would fit the bill in a design. https://power.sager.com/hrpg-600-5-2525734.html
 
  • Like
Reactions: dlerious and cbn

cbn

Lifer
Mar 27, 2009
12,968
221
106
The average capacity drive has been decreasing the amount of 5V power required, but a high end estimate is about 0.6 Amps of 5V power per drive. At 60 drives, you're at 36 Amps of 5V power, or about 200 watts of 5V power (with margin). A high performance drive might use 0.5 Amps of 5V power in 2.5" form factor. The reason they do the above is because commodity power supplies like that are dirt cheap. If you needed a dedicated PSU, there are lots of manufacturers like Meanwell that produce dedicated voltage PSUs that would fit the bill in a design. https://power.sager.com/hrpg-600-5-2525734.html

Thanks for finding that power supply, I think that is needed because the low rpm 2.5" drives don't have any 12v to reduce the amount of amps needed on the 5v. (This to allow them to also work as usb 3.0 drives without power bricks ).

And 120 amps on 5v should also cover the energy usage if all drives were reading/write simultaneously.

The only thing I don't think 120 amps on 5v would cover is the server booting up (180 2.5" drives each spinning up at 1 to 1.2 amps* on 5v simultaneously). However, there is an answer to this problem called staggered spin up (assuming the SATA card has this feature):

*Estimate taken from Start-up current of Seagate Barracuda 2.5".



With that mentioned, the hard drive must also support staggered Spin-up:

https://en.wikipedia.org/wiki/Spin-up

Staggered spin-up
In computers with multiple hard drives, a method called staggered spin-up can be employed to prevent the excessive power-consumption of spin-up, which may result in a power shortage. Power consumption during spin-up is often the highest power draw of all of the different operating states of a hard disk drive. Staggered spin-up typically starts one drive at a time, either waiting for the drive to signal it is ready or allowing a predefined period of time to pass before starting the next drive. If the power supply is able to supply sufficient current to start multiple drives at a time, that, too, is common.

Staggered Spin-up (SSU) and Power-Up In Standby (PUIS) are different features that can help control spin-up of multiple drives within a computer system or a disk subsystem. Both are defined in the ATA Specifications Standards. See Serial ATA for more information.

One feature, called Power-up in standby (PUIS) (also called PM2[1]) is used on some Serial ATA (SATA) and Parallel ATA (sometimes called PATA or IDE) hard disk drives. PUIS requires BIOS and/or driver support to use. When power is applied to the hard disk drive, the drive will not spin-up until a PUIS Spin-Up command is issued. The computer system BIOS or RAID controller must issue the command to tell the drive(s) to spin-up before they can be accessed. PUIS can be enabled by tools such as hdparm for drives which support this feature.

Another feature, called Staggered Spin-up (SSU) is used on most Serial ATA (SATA) hard disk drives. This is more common than Power-Up In Standby (PUIS) because it does not require any special commands to get the drive to spin-up. The drive electronics waits for the SATA Data Phy (Physical I/F) to activate to spin-up the drive. The computer system BIOS and/or RAID controller or RAID driver set can delay and control when the different drives will spin-up.

With Western Digital hard disk drives, Pin 11 of the SATA Power Interface controls whether Staggered Spin-Up (SSU) is enabled or not. Pin 11 is also used as an activity LED connection. When the drive is initially powered on, the drive senses whether Pin 11 is left floating (high or '1' logic state) or grounded (low or '0' logic state). SSU is disabled when Pin 11 is grounded. When disabled, the drive will spin-up as soon as power is applied to it. SSU is enabled when Pin 11 is left floating or driven high (high or '1' logic state). The drive will not spin-up until the SATA Phy Interface becomes active with a connection to a SATA controller or SATA RAID controller. The SATA or SATA RAID controller can control when and how many drives can be spun-up. SSU and PUIS are features that are configured in software or firmware by the manufacturer.

Information from the Fujitsu Serial ATA Interface for Mobile Hard Disk Drives whitepaper:[2]


"Staggered spin-up is a simple mechanism by which the storage subsystem controller can sequence hard disk drive initialization and spin-up. Having this feature not only provides greater reliability, but it allows the system to avoid power surges if all of the HDDs spin up simultaneously during system power up (in a multi-drive environment). Another benefit to having staggered spin-up is the use of more cost-effective power supplies, which prevents power supply damage and system brownouts."

Note that staggered spin-up of disks is a feature of many multi-drive systems using SATA and RAID. It is not typically used on mobile platforms.
 
Last edited:

thecoolnessrune

Diamond Member
Jun 8, 2005
9,673
583
126
You're thinking of consumer 2.5" drives, which are specifically purposed like you mentioned to use only 5V power. Enterprise drives use a combination of 12V and 5V just like their 3.5" counterparts. The reason 5V only drives aren't typically used in high disk arrays is because as you decrease Conversion Voltage, efficiency tends to really take a hit. But, for heavy industry and plentiful pockets, there's always options. Meanwell makes such a PSU if someone ever wanted to turn an entire 120V North American 20 Amp Circuit into 5V power. https://www.arrow.com/en/products/rsp-1500-5/mean-well-enterprises

Keep in mind why that starts reaching into absurdity. To wire something like that up, you'd be talking needing 4/0 (0000) guage wire into a terminal block to distribute. You'd need a carefully laid out wire plan of lower and lower guage wiring (or bus bars) to even approach what you need to go to an individual drive. You start to get massively out of the range of commodity hardware when you're talking $300 power supplies and custom wiring looms.

All I mean is that if you haven't seen something done that seems novel, it's usually because there's a reason :)
 
  • Like
Reactions: FaaR and cbn

cbn

Lifer
Mar 27, 2009
12,968
221
106
You're thinking of consumer 2.5" drives, which are specifically purposed like you mentioned to use only 5V power. Enterprise drives use a combination of 12V and 5V just like their 3.5" counterparts. The reason 5V only drives aren't typically used in high disk arrays is because as you decrease Conversion Voltage, efficiency tends to really take a hit.

The Enterprise capacity 2.5" drives I have seen (this is one of them) are 7200 rpm and use 2.25x to 2.5x more power during write and reads (with greater than 3x the idle) than the 2.5" Seagate Barracuda 5400 rpm I linked in post #12.

So in a case like that I can definitely see splitting the power up between 12v and 5v rather than having 2.5x more amps on the 5v.

But what about a Helium 2.5" 15mm that was only 5400 rpm and used a lot less power than the current capacity 2.5" @ 7200 rpm. In this case (with much lower wattage) does the 2.5" get 12v and 5v? Or just 5v?
 

cbn

Lifer
Mar 27, 2009
12,968
221
106
Thinking about POD 6.0 more, I do have to wonder how much disk activity each of the servers at Back Blaze really gets. This because 60 3.5" hard drives are fed by a total bandwidth of PCIe 2.0 x 6 (ie, 3 x PCIe 2.0 x 2 SATA cards).

So how many amps on the two EVGA 750W PSU(s) 5v rail would 180 2.5" 15mm 5400 rpm HDDs really need at peak usage? (Back Blaze breaks up each file into 20 small pieces called a shard and then those shards are stored in 20 hard drives spread out across 20 Servers.....so we do know that activity is well balanced throughout their data center....but how much activity at peak?)

If the peak activity is rather low with 60 3.5" HDDs per POD, then maybe 180 2.5" HDDs could work (with staggered spin-up for boot) with the existing 2 x 750W PSUs (each with 25 amps on the 5v rail)? (Or more accurately two (or more) SFX PSUs (each with 22 amps on the 5v rail)).

*SFX PSU needed because ATX PSU is too tall for a 3U enclosure.

In fact, they even had a blog about a POD 5.0 Hack where one PSU could be removed when a certain 3.5" hard drive is used (Seagate 4TB, ST4000DM000):

https://www.backblaze.com/blog/storage-pod-5-0-hack/

The design of our overall data storage system spreads out the operational load on any given Storage Pod. This means that at any given time a “majority” of drives in a Storage Pod are idle. The lower power requirements of the Seagate 4TB drive during idle and operation means the total power draw from the 45 hard drives in a Storage Pod does not exceed the specifications of the single PSU on the 5V or the 12V rails.

back-of-1-psu-pod.jpg



With this noted, POD being an open design is getting used by other entities besides Back Blaze....so that is something to think about also.
 
Last edited:

thecoolnessrune

Diamond Member
Jun 8, 2005
9,673
583
126
The Enterprise capacity 2.5" drives I have seen (this is one of them) are 7200 rpm and use 2.25x to 2.5x more power during write and reads (with greater than 3x the idle) than the 2.5" Seagate Barracuda 5400 rpm I linked in post #12.

So in a case like that I can definitely see splitting the power up between 12v and 5v rather than having 2.5x more amps on the 5v.

But what about a Helium 2.5" 15mm that was only 5400 rpm and used a lot less power than the current capacity 2.5" @ 7200 rpm. In this case (with much lower wattage) does the 2.5" get 12v and 5v? Or just 5v?

I wouldn't be certain it's a big difference. The use of Helium is really about getting more platters in there, which, with more rotating mass to put energy into, eats up your savings. The Exos X10 is Helium at 7200 RPM and uses 8.5W Typical in SATA configuration. The Western Digital Red in 10TB only uses 6W Typical at 5400 RPM. Shrinking down the rotating mass nets you even less power savings. With Helium and a slow RPM you'd get low power usage, but less than you'd probably think. I imagine that's why the drives don't really exist right now (the cost / performance / power metrics just don't work out).

I think if you used real-world layouts when you combine port layout, access layout, electricals, thermal room, etc, a single 12TB SMR Helium 3.5" drive beats the density of a stack of many 2.4TB SMR 2.5" drives when the above issues are taken into account.
 

cbn

Lifer
Mar 27, 2009
12,968
221
106
I wouldn't be certain it's a big difference. The use of Helium is really about getting more platters in there, which, with more rotating mass to put energy into, eats up your savings.

Oh yeah, I agree Helium is not about reducing energy consumption much.

Reducing rotational speed is what causes a 2.5" 5400 rpm drive to use less than half the power of a 2.5" 7200 rpm (re: higher velocity results in the power requirement being cubed)

But with regard to a 2.5" x 15mm 5400 rpm Helium drive I am hoping strongly they could go from five platters to seven platters with a small decrease in power similar to the 23% WD got going from five platters to seven (with Helium) while keeping RPM the same:

hgst_hdd_helium_advantage_anandtech.png
 
Last edited:

cbn

Lifer
Mar 27, 2009
12,968
221
106
Here is a cross section of the five platter 3.5" (Air) vs. the seven platter 3.5" (He):

hgst_wd_helium_hdd.jpg


Current 3.5" Helium is eight platters while still keeping 7200 rpm.

I wonder if a further gain in capacity (via reduction in platter thickness) could be had by lowering Helium 3.5" HDD rpm from 7200 to 5900?

Likewise if 2.5" 15mm goes Helium at 5400 rpm, could there by a gain in capacity (via thinner platters) by going 4200 rpm?
 

VirtualLarry

No Lifer
Aug 25, 2001
56,570
10,202
126
Keep the same SATA card arrangement and use port multipliers that turn one SATA into fifteen?

Or should more SATA cards and/or SATA cards with more ports be used and keep port multiplication at 1:5? (Keep in mind no matter approach is taken new backplanes (with the SATA port multipliers) will be needed due to the 2.5" having different drive spacing)
I believe that the limit of five for port multiplier counts is a feature of the SATA protocol. I don't think that you can port-multiply one port into fifteen. Even if you could, wouldn't the throughput be horrible?
 

cbn

Lifer
Mar 27, 2009
12,968
221
106
I believe that the limit of five for port multiplier counts is a feature of the SATA protocol. I don't think that you can port-multiply one port into fifteen. Even if you could, wouldn't the throughput be horrible?

According to this article 1:15 is possible with SATA.

However, you make a good point about throughput if all the drives (or even 1/5 of the 2.5" drives) were using the sata ports at the same time. This especially when you consider that the 12 SATA ports being multiplied come from a mere PCIe 2.0 x 6 (via three PCIe 2.0 x 2 four port cards), not PCIe 2.0 x 12.
 
Last edited:
  • Like
Reactions: VirtualLarry

cbn

Lifer
Mar 27, 2009
12,968
221
106
Regarding the idea of 4200 rpm for 2.5" form factor, this should lower power consumption a bit more than 50% compared to 5400 rpm. This, of course, at the cost of reduced sequential read/write and IOPs.

However, I figure it would also allow 2 x 2.5" 7mm 4200 rpm drives to replace 1 x 2.5" x 15mm 5400 rpm at the ~same 5v power draw.

Maybe even have usb 3.0 external enclosures with 2 x 2.5" 7mm 4200 rpm (in RAID 0 or RAID 1) vs. 1 x 2.5" x 15mm 5400 rpm? (RAID 0 2 x 2.5" 7mm 4200 rpm drives would have greater sequential read/write and IOPs compared to 1 x 2.5" x 15mm 5400 rpm. RAID 1 2 x 2.5" 7mm 4200 rpm would have greater reliability at the cost of capacity and speed compared to 1 x 2.5" x 15mm 5400 rpm)

Or would it even be possible to have a single 1 x 2.5" x 15mm drive composed of 2 x 2.5" x 7mm internally? (ie, internal RAID controller and internal 1:2 "SATA port multiplier").
 
Last edited:

cbn

Lifer
Mar 27, 2009
12,968
221
106
@VirtualLarry,

Regarding the throughput, With each file being broken up (by Back Blaze) into 20 shards (3 of which are for parity) and going to 20 different hard drives in 20 different PODs I have a strong suspicion that random performance of the hard drive will be the limiting factor. This rather than each SATA port, the number of PCIe lanes feeding the SATA card or the bandwidth of the NIC(s).

In fact, looking back on the specs of POD 4.5 I was quite surprised to see the motherboard only had 2 x 1Gb Ethernet (<---scroll down the page after clicking the link to see the POD 5.0 vs. POD 4.5 hardware comparison). That means even when the two NICs were teamed the maximum input was only 250 MB/s to cover the needs of 45 3.5" hard drives. This compared to the current arrangement of 2 x 10Gb Ethernet which when teamed together provides 2500 MB/s to the 3 x PCIe 2.0x2 SATA cards.

P.S. The Blog article for POD 4.5 is dated March 5th, 2015. So it wasn't that long ago for 2 x 1Gb Ethernet.
 
Last edited:

Charlie22911

Senior member
Mar 19, 2005
614
231
116
@VirtualLarry,

Regarding the throughput, With each file being broken up (by Back Blaze) into 20 shards (3 of which are for parity) and going to 20 different hard drives in 20 different PODs I have a strong suspicion that random performance of the hard drive will be the limiting factor. This rather than each SATA port, the number of PCIe lanes feeding the SATA card or the bandwidth of the NIC(s).

In fact, looking back on the specs of POD 4.5 I was quite surprised to see the motherboard only had 2 x 1Gb Ethernet (<---scroll down the page after clicking the link to see the POD 5.0 vs. POD 4.5 hardware comparison). That means even when the two NICs were teamed the maximum input was only 250 MB/s to cover the needs of 45 3.5" hard drives. This compared to the current arrangement of 2 x 10Gb Ethernet which when teamed together provides 2500 MB/s to the 3 x PCIe 2.0x2 SATA cards.

P.S. The Blog article for POD 4.5 is dated March 5th, 2015. So it wasn't that long ago for 2 x 1Gb Ethernet.

Apples to oranges I know, but I'm running a dual parity unRAID box and my experience is that random performance is the bottleneck. I imagine having a greater number of disks for parity operations helps to a degree versus the unRAID implementation though.
 

BonzaiDuck

Lifer
Jun 30, 2004
16,122
1,738
126
I have a casual interest in this 2.5" drive question that is growing, partly as result of the discussion here.

Since SSD technology and NVME, those types of disks become the hub for the OS and programs. HDDs still come in large capacities and offer different dimensions of reliability: for instance, your data should be safe on an HDD that has not been powered up for a year or longer. Risks of total data loss from hardware failure are different for the two types of media.

Therefore, I've been using 2.5" laptop drives for media, data and backup in my desktop systems. I have not had any problem with this at all. The drives are lighter than 3.5" units and consume less power. If the system benefits from faster access to those disks which they don't natively provide, I can cache them to RAM as might work well for certain tasks and file sizes. Or, I can cache them to a persistent SSD-cache using a small SSD or NVME SSD. For usage such as media (movie files) and backup, there shouldn't be much of an issue about performance, or the need to for a caching solution. Either way, I don't fret much over the RPM spec of 5,400 vs 7,200 or higher.

Certainly, I'll be interested to see what happens with 2.5" drives -- laptop or enterprise -- in the near future. That includes filling them with He, the platter limit, and other factors.
 

cbn

Lifer
Mar 27, 2009
12,968
221
106
https://patents.google.com/patent/US6546499B1/en

^^^^ This patent is called "Redundant array of inexpensive platters (RAIP)".

The present invention relates in general to using data management and storage techniques and concepts from Redundant Array of Independent Disks (RAID) technology and incorporating these techniques and concepts into a single disk drive and in particular to providing and using a redundant array of inexpensive platters (RAIP) within a single disk drive.

The RAID technology provides excellent solutions for storage and high performance access of data. However, the use of multiple disk drives, at times and instances, may be cost prohibitive, expensive, and infeasible in implementing the RAID methodology for deriving the advantages therefrom for desired applications and purposes. Thus, it would be highly desired at these times and instances to incorporate the RAID concepts and techniques into a single disk drive, particularly for providing the cost advantages of using less disk drives.

It was filed back in 1999, but I'm thinking with Helium this could be a lot more interesting.

For example, lets say that Internal RAID-0 is possible, how fast do the platters really need to be spinning to saturate SATA 6 Gbps?

SAS 24 Gbps?

U.2? (PCIe 4.0, PCIe 5.0, etc)

Power consumption at these speeds?
 
Last edited: