Seagate brings multi-Actuator Tech soon

[Elixer]

Seagate unveiled today its new Multi Actuator technology, a breakthrough that can double the data performance of its future-generation hard drives in hyperscale data centers. As higher areal densities on future hard drives put downward pressure on performance, Seagate’s Multi Actuator technology will more than offset these pressures. That means customers with data-intensive applications will continue to enjoy the highest levels of hard drive performance, while they simultaneously keep up with the need to manage vast, ever-increasing quantities of data. Seagate’s Multi Actuator technology is in development to be deployed on products in the near future.

https://blog.seagate.com/technology..._flagship3_messaging;JXyixoGaROyvrtBYgUlcBw==

In its first generation, Seagate’s Multi Actuator technology will equip hard drives with dual actuators (two actuators). With two actuators operating on a single pivot point, each actuator will control half of the drive’s arms. Half the drive’s recording heads will operate together as a unit, while the other half will operate independently as a separate unit. This enables a hard drive to double its performance while maintaining the same capacity as that of a single actuator drive.

This could be very interesting! 2x the chance of a head crash!

 

[misuspita]

I remember as a teenager looking at all those platters and thinking why nobody makes every single head move independently.

Sooo... Why? In a 5-6 platter hard-drive one head is reading at one time, and the other sit there doing o thing.

 

[Insert_Nickname]

This could be very interesting! 2x the chance of a head crash!

Everything else being equal, this should help a fair bit with mixed R/W workloads. Lets see if they can pull it of reliably... (being Seagate, you could have a few doubts)

I remember as a teenager looking at all those platters and thinking why nobody makes every single head move independently.

Sooo... Why? In a 5-6 platter hard-drive one head is reading at one time, and the other sit there doing o thing.

I suspect independant actuators would make the actuator assembly a little complicated. Not to mention the additional management overhead, and processing requirements.

But yeah. Its a cool idea.

 

[cbn]

I know this tech is aimed at machines like this one.

2 x24 Core (2.1 GHz) Intel ® Xeon ®8160
8 x 32GB DDR4 2666 MHz Memory expandable to 1.5TB
12 x 10TB 7,200 RPM High Capacity SAS Drives
2 x 150GB M.2 SATA SSD Drives
2 x QDR 40Gb/sec InfiniBand Ports
1 x Dual port InfiniBand QDR CX3 (40 Gb/sec) PCIe HCA
1 x Built-in RJ45 1 Gigabit Ethernet por

But I do wonder if this would also come to 10,000 rpm SAS (2.5")?

 

[Elixer]

I know this tech is aimed at machines like this one.



But I do wonder if this would also come to 10,000 rpm SAS (2.5")?

I am betting it will be for enterprise first.

 

[cbn]

One reason I am asking is because I see (in some instances) multiple small capacity 10K SAS drives are used rather than one larger 10K SAS. Example would be Oracle Exadata X7-2 machine (specs on this page.)

Notice the database server has 4 x 600GB 10K SAS drives (ie, 600GB platters and 4 actuators) rather than one 2.4TB 10K SAS drive (ie, four 600GB platters with one actuator).

P.S. There is also another Oracle Exadata machine called X7-8 which doesn't use SAS drives for the database server. (It is all flash, specs here)

 

[cbn]

Here are the specs for the current Seagate 512 emulation/4K native 10,000 rpm SAS drives:

I wonder if moving these 10,000 rpm 2.5" platters into a 3.5" housing would give Seagate a greater opportunity to add actuators compared to leaving them in a 2.5" housing? (Eg, 3.5" housing with 8 actuators for eight 2.5" platters vs. 2.5" housing with 2 actuators for four 2.5" platters)

 

[cbn]

Here is an article where the Register asked Seagate and some IT experts about the possibility of a dual Pillar multi-actuator drive:

https://www.theregister.co.uk/2018/01/04/doubling_up_disk_drive_actuator_pillars/

So unlike the current idea of having 2 (or more) actuator arms on a single pivot there would an additional pivot added to the drive enclosure (adding another 2 (or more) actuator arms). See example below (from this 2009 article) of a two pivot point drive:

P.S. Using the following WD Velociraptor (2.5" platter in a 3.5" Icepack enclosure) as reference point I reckon Seagate could get a ~3" platter in between 2 actuator pillars sitting catty corner in a 3.5" enclosure.

 

[cbn]

A patent from Seagate with two actuator arms on single pivot:

https://www.google.com/patents/US6121742

In a computer disc drive system having a plurality of voice coil motors for independently actuating a plurality of actuator arms or E-blocks, a single voice coil motor servo control circuit controls the plurality of voice coil motors. In a preferred embodiment, a single digital servo processor dynamically allocates servo bandwidth to the plurality of voice coil motors according to the present activity of each voice coil motor. In a preferred embodiment, the single digital servo processor allocates the greatest amount of bandwidth to a voice coil motor that is presently positioning an associated read/write head for writing, a lesser amount of bandwidth to a voice coil motor that is presently positioning a read/write head for reading, and the least amount of bandwidth to a voice coil motor that is presently seeking or idle.

 

[cbn]

Looking at the abstract from the Seagate patent in the previous post I wonder how much of issue servo bandwidth would be using a modern digital servo processor with something like 8 actuators?

The abstract mentions positioning a head for writing uses the most bandwidth followed by positioning a head for reading.

In a preferred embodiment, the single digital servo processor allocates the greatest amount of bandwidth to a voice coil motor that is presently positioning an associated read/write head for writing, a lesser amount of bandwidth to a voice coil motor that is presently positioning a read/write head for reading, and the least amount of bandwidth to a voice coil motor that is presently seeking or idle.

If it is an issue maybe the "high actuator count design" could be based on maxing out read and then using a buffer or cache for completing some of the writes. (eg, four actuators write half of the file first while the second half of the file is cached for the second set of the four actuators to do their writing next. Basically with this approach sequential write is halved compared to the sequential read....but the sequential read would be higher than it would have been without the use of the cache).

If running out of buffer or cache is an issue maybe the drive could slow down RPM during writing if this helps conserve servo bandwidth so all actuators can be used positioning the heads for simultaneously writing a large file?

P.S. I know IOPS are important (and a lot people believe this tech is mostly for IOPs), but as mentioned in this greybeards podcast even Western Digital admits average block sizes are growing in this era of rich data. (reference 28.30 (and 28:40 ) to 29:26 in the pod cast. So throughput, much more than IOPs (<---This mentioned at 28:47 to 28:50 by one of the greybeards)). Of course, RAID rebuild times is another thing to consider (boosting Sequential helps this) and I wonder how much it helps file transfer within a server system like this one which uses the Exalytics in-memory machine*).

*I've never used Exalytics myself, but in this post it is described as smart read cache which suggests parts of the database to be replicated into its memory and storage. (So the data does need to be copied into the Exalytics machine.)

 

[WhoBeDaPlaya]

What's old is new again.
I think I have one around that actually still works!

 

[aigomorla]

someone tell seagate this is the wrong way to R0 something.... its best if we have 2 complete drives, and not 2 head actuators...

 

[Insert_Nickname]

What's old is new again.
I think I have one around that actually still works!

At least it didn't sound like a turbine impression, unlike some of the other 52X/56X drives.