Tuesday, June 25th 2024
Toshiba Announces MG10-D Series of Enterprise HDDs with Capacities up to 10TB
Toshiba Electronics Europe GmbH (Toshiba) announces the release of its MG10-D Series, a family of air-filled conventional magnetic recording (CMR) HDDs supporting SAS and SATA interfaces and capacities of up to 10 TB. Crafted with precision engineering and over 50 years of Toshiba experience, the MG10-D Series delivers improved performance and power efficiency over prior generations. With sanitize instant erase (SIE) and self-encrypting drive (SED) options, valuable data is safeguarded by a storage solution known for its robust performance and unwavering dependability.
Built for the increasing application demands of enterprise server and storage solutions, the MG10-D Series delivers a new level of performance. For example, compared with the previous model, the new 10 TB MG10ADA10TE provides an approximately 13% better maximum sustained transfer speed of 268MiB/s and doubles the cache buffer size to 512MiB. It also reduces power consumption in active idle mode by approximately 21%, to 5.74 W. Architected to deliver improved total cost of ownership (TCO), the new MG10-D Series fits seamlessly into a wide variety of business-critical applications, such as email, data analytics, data retention, and surveillance.
"Toshiba's MG10-D Series delivers exceptional performance to meet the demands of growing business critical applications. The new cutting-edge design of the MG10-D Series is engineered for sustainable enterprise environments and fits seamlessly into existing infrastructure reducing TCO," said Larry Martinez-Palomo, Vice President, Head of Storage Products Division at Toshiba.
The MG10-D Series is a 5-disk CMR standard 3.5-inch, 7200 RPM air-filled platform. Available capacities are 2 TB, 4 TB, 6 TB, 8 TB, and 10 TB for both SAS and SATA. SATA is also available in a 1 TB drive. The series supports 6 Gb/s SATA or 12 Gb/s SAS interface options in Advanced format 512e and 4Kn. A 512n option is available on the 1 TB, 2 TB, and 4 TB offerings to support legacy systems with native 512 byte block sizes. Designed for 24x7 enterprise reliability, the MG10-D Series has a workload rating of 550 TB, an AFR of 0.44% and an MTTF/MTBF of 2M hours.
The MG10-D Series will be available in CQ3.
Source:
Toshiba
Built for the increasing application demands of enterprise server and storage solutions, the MG10-D Series delivers a new level of performance. For example, compared with the previous model, the new 10 TB MG10ADA10TE provides an approximately 13% better maximum sustained transfer speed of 268MiB/s and doubles the cache buffer size to 512MiB. It also reduces power consumption in active idle mode by approximately 21%, to 5.74 W. Architected to deliver improved total cost of ownership (TCO), the new MG10-D Series fits seamlessly into a wide variety of business-critical applications, such as email, data analytics, data retention, and surveillance.
The MG10-D Series is a 5-disk CMR standard 3.5-inch, 7200 RPM air-filled platform. Available capacities are 2 TB, 4 TB, 6 TB, 8 TB, and 10 TB for both SAS and SATA. SATA is also available in a 1 TB drive. The series supports 6 Gb/s SATA or 12 Gb/s SAS interface options in Advanced format 512e and 4Kn. A 512n option is available on the 1 TB, 2 TB, and 4 TB offerings to support legacy systems with native 512 byte block sizes. Designed for 24x7 enterprise reliability, the MG10-D Series has a workload rating of 550 TB, an AFR of 0.44% and an MTTF/MTBF of 2M hours.
The MG10-D Series will be available in CQ3.
21 Comments on Toshiba Announces MG10-D Series of Enterprise HDDs with Capacities up to 10TB
At the right price point, this is a very good buy IMO. But that's the big question, what price is this coming out at? With 18TB drives in the $250 price (albeit consumer), I'd say $200ish for a 10TB Enterprise would be fair.
I do run my own NAS at home and I don't see much need to go expensive Helium or HAMR or go much beyond 8TB. After all, a NAS is going to be composed of 4x hard drives (redundancy + speed), so 4x 8TB drives (effective 16TB space, 2 drive redundancy) is far more useful to me than 2x 16TB drives (which only has 1x drive redundancy). And I'd pay a premium for the extra drives. Its hard to put a number on reliability after all.
Hard Drive users today have to be mostly NAS-boxes, or maybe datacenter-scale NAS. And I expect most home NAS or even small-office commercial users would want ~4x HDDs at a minimum, and likely don't need much more space than 16TBs (total across 4x hard drives of 8TB each).
for the most part in my career as a storage architect and just working on storage when I was in other roles; most companies are moving from raid 5 and 6 as a whole and I also will never implement it. They are actually not desired as storage prices fall and density goes up, because the chance of a URE increases and the odds you can rebuild that much capacity in those raid tiers falls dramatically.
for the price it makes much more sense to balance r/w performance and simply add more disks to the spans if you needed added hot spare redundancy.
You really don't want RAID5, or any other architecture which can only withstand 1x error. You need to withstand 1x error, and then allow for a rebuild (where the rebuild itself can withstand a 2nd error). After all, its the rebuild where the 2nd error is often discovered.
Not that I do RAID anymore, but instead ZFS mirroring / parity / etc. etc., or Windows Storage Spaces if I ever plan to use Windows. Modern software has gotten much better and should be the default choice than the older RAID drivers. (And modern CPUs are so fast that the extra CPU-hit on all those I/O calculations is fine. Except on Windows: parity IIRC was still broken and awful performance-wise last time I checked)
In any case, 4x hard drives are needed for a good long-term solution. Be it RAID10, RAID6, ZFS Parity with 2 drive redundancy, or Windows Storage Spaces with striped+mirroring enabled. 4x drives maps to a large number of software and hardware solutions that its the best starting point.
As far as I know WD only ever got up to 1.67TB platters, anything above that used SMR. So if Toshiba got 2TB platter CMR drives, in 4x platters for the 8TB drive, then they'd beat WDs offerings.
On the other hand Toshiba only seems to have 7200rpm drives so probably not.
WD air filled drives still cap at 1.67TB platters as far as I know.
Since it requires more advanced tech to reach 20TB and beyond, there are questions in my mind if its "worth it", especially for the typical home consumer. Having a collection of cheaper, traditional hard drives built off of a lower-technology platform could be interesting moving forward.
I'm looking at 100-500TB solutions a few times a year and right now a lot of providers will try to convince you to go all-flash which completely overdelivers on performance and completely underdelivers on price.
Despite all the major players trying to get away from hybrid (ie, mechanical disks - because almost nobody makes multi-drive solutions without a flash cache these days) the issue is that AF arrays are still too damn expensive for a lot of situations and if a hybrid array can delivery 250TB of storage with adequate IOPS for 1/3rd the cost of an AF solution, then that's what we'll take.
Personally, I'm interested in taking EOL production storage that was sold 6+ years ago with 4TB drives and replacing the drives to use it as tier-2 storage that's not mission-critical. Highly-compatible, basic, guaranteed-to-work disks still have a large market, even if they're not at the bleeding edge of cost/TB...
If you are having a resiliency strategy (RAID6, RAID10, or similar), do you seriously expect to fill up 40TB of data in any time in the next decade?
These amounts of TB are basically infinite to me either way. 10TB is similar to 20TB or 40TB. I can safely say that I simply do not generate that kind of data size even across years of my life.
The only net gain I can do is resiliency and convenience. NAS for iSCSI and network storage features.... And then guarantees about needed to break 2+ hard drives before I lose data.
Its not a "big" chance of failure, but its often enough that you should plan for it. Especially because we only know what the "true" chance of failure is 4 or 5 years from now after a hard drive model has been deployed and used for years in practice.
I'm not disputing your figures, just surprised that I have so many spinning disks (1000 or so) that result in a drive failure 2-5x a year for a typical disk life under support contract of around 6 years. (~2.5% of drives fail during their whole lifespan, which is a 0.4% annual failure rate.
Perhaps all the stats we ever see for HDD failures are consumer-driver stats, or total manufacturer RMA rates that are weighted-up by consumer stats. I know Backblaze publish annually and they had a history of buying the cheapest external drives they could find at retail and shucking them, though I'm not sure if they still do that....
Any other price except this is, it's an callous overpricing.
You have to consider that soon we are going to have 8TB, 12TB and 16TB QLC and PLC SSD Drives, that are going to blow those out of water, for almost similar prices in the future.
You've basically taken the log-plot of the following Simpson's joke and are unironically using it to predict the future.
At ~$320ish I pulled the trigger on a 16TB MG08 series Toshiba with that giganto 512MB cache and for the past few years it's been hosting my iSCSI volumes.
This year it went from cold spare to full time 24/7 duty and it's great. The MG10 series drives aren't targeted for my use case but if sub-$130ish at 10TB, that's a steal.
Writes are always gonna suck because iSCSI but so far the reads are really nice. I probably made the best choice here.