In our office 5 out 6 Vertex 2 failed showing that SSDs are not ready for the enterprise. BTW, they were in completely different systems with AMD or Intel chipsets, in desktops or laptops. Very dissapointed...
HAHA you bought Vertex drives for your office? The guy behind that decision should be fired, along with you for thinking OCZ represents the SSD market.
Actually, enterprise simply adapts to the current issues with SSD and uses them intelligently. Evidence to that was the SNW conference this month where SSD storage in enterprise was one of the 3 main buzz topics, and pretty much every enterprise storage vendor had a solution involving SSD drives, either for tiered storage for caching, or for analysis scratch space. Huge performance benefits, virtually no data loss risk.
This is great news. Sandforce clearly has some brilliant engineers, but they need the QA resources to back them. LSI provides just that; this should be good for everyone.
LSI doesn't have to sell to the consumer market. If they enforce enterprise class validation and testing on Sandforce controllers, why couldn't other vendors buy the controllers to use in their own SSD's?
LSI has been making storage silicon for years and understands very well the importance of validation on a wide variety of platforms so this should be a huge win for SF.
LSI and Intel are also quite chummy so one has to wonder if this truly paves way for the mythical Intel 520 with a SF controller.
OCZ is in for a world of hurt trying to compete with their home-brew controller now. They don't have the resources, the experience, the support, or the accountability to their customers to be able to succeed on their own silicon. I expect them to continue to trip over their own feet and fail miserably.
I don't know if I would call Everest "home-brew". Indilinx did quite well with its first controller, and that was without the resources of OCZ. Putting those resources behind the same team could make a great product, and I look forward to seeing what it is able to do.
Isn't everyone forgetting Samsung makes controllers, and they happen to sell in the consumer sector as well? What about Marvel, the controller company that Intel chose over their own chip? Lets not pretend OCZ and Intel are the only SSD manufacturers. They are certainly not the only supplier of controllers.
They are up against those like Intel, Marvell, Samsung, Toshiba, and now LSI -- those who have been doing semi design work for quite some time and have very good validation processes.
I just don't see OCZ having the resources to match the above as eventually it all comes down to how fast you can iterate designs, and how fast you can validate those designs and get them to market. This all comes down to people and talent and I don't see OCZ being able to make the same investments as the above in these areas - who also have a huge head start.
OK, with LSI and their expertise in RAID controllers, along with the enterprise level controllers Sandforce produces, I forsee one product that could change servers forever: an LSI single chip concoction of direct PCIe to SLC flash memory for a totally integrated solid state booting system taking up no drive bays or card slots, just some real estate on the motherboard. I also see a competitor for the Revo Drive coming from LSI that could potentially be cheaper and better performing, with a write cache that would protect performance from the slow random writes we see in SSD RAIDs. (I happen to have an LSI 9750-8i RAID controller with 2 Vertex 2 drives, along with other drives. the write caching makes a HUGE performance difference.)
oh, yeah, forgot about that thing. However, with LSI owning Sandforce, they'll be able to get the SSD controllers cheaper and could potentially make this thing cheaper. I'd like that.
On the one hand I have a Vertex 3, so I'm the first guy to sign up for more betterer QA. Then again, I was hoping that would happen through cooperation with for example Intel. Time will tell I suppose.
I'm going to play devil's advocate for a moment and say that the enterprise markets could really benefit from reliable SSD's. In my 8-10 years of working in the enterprise in IT, one of the biggest agents of "age" in a machine - laptops and USDT's in particular - is the hard drive. Older hard drives are, generally, considerably slower than those installed in newer families of machines bought by businesses (perp. storage, greater platter densities, better controllers, etc.). The user experience on traditional HDD's degrades to such extents that many users in many companies would just as soon have their desktop or laptop replaced, even though the only real problem with it is an aging HDD.
Reliable SSD's in my particular field would give at LEAST another 2 years out of most workstation types. Our laptops for our mobile force would be faster with better battery life, our desktops would be faster with considerably lower power envelopes and noise. And now that SSD's are building encryption capabilities directly into the controllers, compliance-laden industries like mine (Healthcare) can start giving them serious consideration. It just comes down to reliability and up-front cost, and any company that can bring SandForce performance reliably and cheaply to my field not only brings all the benefits listed above to my field, but makes MY job considerably more entertaining, too! ;D
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
27 Comments
Back to Article
Beenthere - Wednesday, October 26, 2011 - link
Enterprise isn't going to tolerate the SSD compatibility and reliability issues that the consumer segment is enduring.chbarg - Wednesday, October 26, 2011 - link
In our office 5 out 6 Vertex 2 failed showing that SSDs are not ready for the enterprise.BTW, they were in completely different systems with AMD or Intel chipsets, in desktops or laptops.
Very dissapointed...
sanguy - Wednesday, October 26, 2011 - link
And we have literally 100's of Intel 310 and 320 series SSD's in laptops, desktops, embedded systems, and servers and have ZERO failures.....Very happy.
Iketh - Wednesday, October 26, 2011 - link
HAHA you bought Vertex drives for your office? The guy behind that decision should be fired, along with you for thinking OCZ represents the SSD market.The only option right now is Intel.
radium69 - Thursday, October 27, 2011 - link
And crucial ;)scottjames_12 - Thursday, October 27, 2011 - link
Marvell would probably be more accurate - Crucial/Micron make NAND mainly, and drives using Marvell's controllers, if I'm not mistaken.radium69 - Thursday, October 27, 2011 - link
Absolutely right!scottjames_12 - Thursday, October 27, 2011 - link
And Samsung ;)A5 - Thursday, October 27, 2011 - link
This.Blackened144 - Thursday, October 27, 2011 - link
We tried the same thing by purchasing 20 1TB OCZ Colossus drives.. We RMA'd about 10 of them before we decided to stop testing..RMSe17 - Thursday, October 27, 2011 - link
Actually, enterprise simply adapts to the current issues with SSD and uses them intelligently. Evidence to that was the SNW conference this month where SSD storage in enterprise was one of the 3 main buzz topics, and pretty much every enterprise storage vendor had a solution involving SSD drives, either for tiered storage for caching, or for analysis scratch space. Huge performance benefits, virtually no data loss risk.RMSe17
bigi - Thursday, October 27, 2011 - link
Enterprise can pay lots more to get reliable product.Smart consumers who buy good SSDs don't have to endure anything.
dcollins - Wednesday, October 26, 2011 - link
This is great news. Sandforce clearly has some brilliant engineers, but they need the QA resources to back them. LSI provides just that; this should be good for everyone.B3an - Wednesday, October 26, 2011 - link
Not if LSI only make the drives for the enterprise market.Taft12 - Wednesday, October 26, 2011 - link
... I can't see LSI having any interest in the consumer market when they're used to enterprise margins.It's a sad day when much of the consumer SSD market now depends on OCZ/Indilinx.
Etsp - Thursday, October 27, 2011 - link
LSI doesn't have to sell to the consumer market. If they enforce enterprise class validation and testing on Sandforce controllers, why couldn't other vendors buy the controllers to use in their own SSD's?sanguy - Wednesday, October 26, 2011 - link
LSI has been making storage silicon for years and understands very well the importance of validation on a wide variety of platforms so this should be a huge win for SF.LSI and Intel are also quite chummy so one has to wonder if this truly paves way for the mythical Intel 520 with a SF controller.
OCZ is in for a world of hurt trying to compete with their home-brew controller now. They don't have the resources, the experience, the support, or the accountability to their customers to be able to succeed on their own silicon. I expect them to continue to trip over their own feet and fail miserably.
ICBM - Wednesday, October 26, 2011 - link
I don't know if I would call Everest "home-brew". Indilinx did quite well with its first controller, and that was without the resources of OCZ. Putting those resources behind the same team could make a great product, and I look forward to seeing what it is able to do.Isn't everyone forgetting Samsung makes controllers, and they happen to sell in the consumer sector as well? What about Marvel, the controller company that Intel chose over their own chip? Lets not pretend OCZ and Intel are the only SSD manufacturers. They are certainly not the only supplier of controllers.
sanguy - Wednesday, October 26, 2011 - link
They are up against those like Intel, Marvell, Samsung, Toshiba, and now LSI -- those who have been doing semi design work for quite some time and have very good validation processes.I just don't see OCZ having the resources to match the above as eventually it all comes down to how fast you can iterate designs, and how fast you can validate those designs and get them to market. This all comes down to people and talent and I don't see OCZ being able to make the same investments as the above in these areas - who also have a huge head start.
dgingeri - Wednesday, October 26, 2011 - link
OK, with LSI and their expertise in RAID controllers, along with the enterprise level controllers Sandforce produces, I forsee one product that could change servers forever: an LSI single chip concoction of direct PCIe to SLC flash memory for a totally integrated solid state booting system taking up no drive bays or card slots, just some real estate on the motherboard. I also see a competitor for the Revo Drive coming from LSI that could potentially be cheaper and better performing, with a write cache that would protect performance from the slow random writes we see in SSD RAIDs. (I happen to have an LSI 9750-8i RAID controller with 2 Vertex 2 drives, along with other drives. the write caching makes a HUGE performance difference.)MGSsancho - Thursday, October 27, 2011 - link
http://www.lsi.com/products/storagecomponents/Page...dgingeri - Thursday, October 27, 2011 - link
oh, yeah, forgot about that thing. However, with LSI owning Sandforce, they'll be able to get the SSD controllers cheaper and could potentially make this thing cheaper. I'd like that.A5 - Thursday, October 27, 2011 - link
You used "LSI" and "cheaper" in the same sentence. Nice joke, had me going there for a second :-PTroff - Wednesday, October 26, 2011 - link
On the one hand I have a Vertex 3, so I'm the first guy to sign up for more betterer QA. Then again, I was hoping that would happen through cooperation with for example Intel. Time will tell I suppose.farsawoos - Thursday, October 27, 2011 - link
I'm going to play devil's advocate for a moment and say that the enterprise markets could really benefit from reliable SSD's. In my 8-10 years of working in the enterprise in IT, one of the biggest agents of "age" in a machine - laptops and USDT's in particular - is the hard drive. Older hard drives are, generally, considerably slower than those installed in newer families of machines bought by businesses (perp. storage, greater platter densities, better controllers, etc.). The user experience on traditional HDD's degrades to such extents that many users in many companies would just as soon have their desktop or laptop replaced, even though the only real problem with it is an aging HDD.Reliable SSD's in my particular field would give at LEAST another 2 years out of most workstation types. Our laptops for our mobile force would be faster with better battery life, our desktops would be faster with considerably lower power envelopes and noise. And now that SSD's are building encryption capabilities directly into the controllers, compliance-laden industries like mine (Healthcare) can start giving them serious consideration. It just comes down to reliability and up-front cost, and any company that can bring SandForce performance reliably and cheaply to my field not only brings all the benefits listed above to my field, but makes MY job considerably more entertaining, too! ;D
Just my $.02
NandFlashGuy - Friday, October 28, 2011 - link
Intel has "eaten their own dogfood", deploying SSDs across 55% of the workforce (as of end of 2010):http://www.intel.com/content/dam/doc/white-paper/i...
They did see the tocal cost of ownership benefit of SSDs based on a 90% reduction in Annual Fail Rate.
zhangqq - Monday, October 31, 2011 - link
http://ygn.me/bTf7p