Intel announces a new line of SSDs aimed at data centers

[Bouzoo]

Similar to its big brothers, the S3510 features Intel's second generation SATA 6Gbps controller that was first introduced in the high capacity S3500 models late last year. Intel has remained quiet about the specifics of the second generation controller (and the SATA 6Gbps controller as a whole), but we do know that it adds support for larger capacities, which suggests the internal caches and DRAM controller could be larger

 

The most significant change in the S3510 is the NAND. The S3510 switches to IMFT's latest 16nm 128Gbit MLC NAND node, which is a rather surprising move given that all Intel's client SSDs are still utilizing 20nm NAND. The reason lies behind the fact that Intel didn't invest in IMFT's 16nm node, meaning that Micron produces and owns all 16nm NAND output. Intel and Micron reconsider the partnership and investments for each generation separately and for 16nm Intel decided not to invest -- likely because Intel's focus is in the enterprise nowadays and 16nm is more geared towards the client market given its lower endurance, and Intel also wanted to concentrate more heavily in the companies' upcoming 3D NAND.

 

That said, Intel and Micron do have strong supply agreements in place, which gives Intel access to Micron's 16nm NAND despite not investing in its development and production. I suspect the use of 16nm NAND is why the S3510 wasn't launched alongside the S3610 and S3710 earlier this year because validating a new NAND node is time consuming and might be that the 16nm node wasn't even mature enough for the enterprise back then. In any case, the S3510 is the first enterprise SSD to utilize sub-19nm NAND, which is a respectable achievement on its own already. 

 

 

 

 

Seems like 480GB is the sweet spot right?

 

On the performance side, the S3510 provides slightly better random write performance at larger capacities than its predecessor (you can find the S3500 specs here), but other than that the S3510 is a very close match with the S3500. Typical to enterprise SSDs, the S3510 features AES-256 hardware and full power loss protection that protects all data, including in-flight user writes, from sudden power losses. 

 

 

The price is still unknown. Remember, this are not your average consumer SSDs, these are made for writing like crazy all the time.

 

 

Sources:

http://www.anandtech.com/show/9226/intel-releases-ssd-dc-s3510

 

http://www.tweaktown.com/news/45041/intel-announces-new-dc-s3510-ssds-aimed-data-center-market/index.html

[SirRoderick]

Do data centers use SSDs a lot? In my head they still rely mostly on HDD for reasons of space and reliability.

[simson0606]

Do data centers use SSDs a lot? In my head they still rely mostly on HDD for reasons of space and reliability.

some databases rely on the speed of a ssd

[Guest]

Do data centers use SSDs a lot? In my head they still rely mostly on HDD for reasons of space and reliability.

Absolutely. Almost all servers will have an SSD in some form installed. Whether it be the OS host, a storage device for general use or a cache for a RAID card.

[euniqe]

Yay.

 

The closer we come to the death of the mechanical hard-drive the happier I will be. 

[patrickjp93]

Yay.

 

The closer we come to the death of the mechanical hard-drive the happier I will be. 

That won't ever happen. Just as HDD can't supplant tape for archival purposes, neither can SSD supplant HDD for the same purposes. Spinning rust can go up to 20TB using HAMR technology, and more innovations are on their way. Meanwhile SSDs are still beholden to Moore's Law and the collapse of Dennard's Law.

[Bonn]

We use tiered storage in hybrid SANS in the enterprise. These are SATA drives, so immediately out for anyone serious. SAS is the only interface unless its PCI-e cards with mirrored data to another server or DC, or using 2PC transactions. 

[patrickjp93]

We use tiered storage in hybrid SANS in the enterprise. These are SATA drives, so immediately out for anyone serious. SAS is the only interface unless its PCI-e cards with mirrored data to another server or DC, or using 2PC transactions. 

Tell that to Google. Google uses no enterprise-grade equipment in any of its data centers because the failure rates are basically equivalent and the costs are terribly higher (Dr. Keith Fricken, Google DBA and Miami University Professor Emeritus).

[Bonn]

Tell that to Google. Google uses no enterprise-grade equipment in any of its data centers because the failure rates are basically equivalent and the costs are terribly higher (Dr. Keith Fricken, Google DBA and Miami University Professor Emeritus).

Google is very, very different. You cannot compare google, facebook etc they do their own shit with the Open compute project and the likes. 

[patrickjp93]

Google is very, very different. You cannot compare google, facebook etc they do their own shit with the Open compute project and the likes. 

Oh please, Google's cost structure is so much better it should be driving the entire industry to question Dell and Oracle about their exorbitant prices, much less the costs for enterprise drives. And in fact they're directly comparable with OpenCompute. They donate enough compute time to both projects they practically are the projects.

[Bonn]

Oh please, Google's cost structure is so much better it should be driving the entire industry to question Dell and Oracle about their exorbitant prices, much less the costs for enterprise drives. And in fact they're directly comparable with OpenCompute. They donate enough compute time to both projects they practically are the projects.

Web content and search engines are a different story, my point above, the size of google will use their own methods do things, but their platforms are built from the ground up and split over multiple data centers and compute nodes, blah blah. I have this same argument with a friend who works for Facebook who's always recommending stuff, Google and Facebook solutions are NOT viable to everybody. 

I can go whack our corporate hub platforms and bring their storage to it's knees in seconds.

Learn the requirements for each, don't just assume what google do is correct, it's a niche tech not applicable to all. 

[patrickjp93]

Web content and search engines are a different story, my point above, the size of google will use their own methods do things, but their platforms are built from the ground up and split over multiple data centers and compute nodes, blah blah. I have this same argument with a friend who works for Facebook who's always recommending stuff, Google and Facebook solutions are NOT viable to everybody. 

I can go whack our corporate hub platforms and bring their storage to it's knees in seconds.

Learn the requirements for each, don't just assume what google do is correct, it's a niche tech not applicable to all. 

It's applicable to everyone. If you can replace one small set of expensive, specialized equipment with multiple sets of more general equipment which maintains 95+% of the performance, and you're not a nanosecond-critical firm like stock traders, you should replace your infrastructure if the cost structure is sufficiently in favor of the change. It's simple economics. What sort of infrastructure do you run and for what purposes?

[Vanderburg]

I work for a game server provider and we use lots of enterprise Intel SSDs, all over the place (MySQL on webhosting, HDD caching on game server machines, as an add-on for VDS/VPSs and as a standard configuration for many dedicated machines).