backtop


Print 21 comment(s) - last by winterspan.. on May 22 at 2:00 AM

OCZ's highly anticipated new high-end SSD is targeted at enthusiasts and the enterprise market

OCZ Technology has become a major player in the enthusiast storage market, thanks to their strong Solid State Drive portfolio. The company has been very successful with its Vertex series of SSDs, but strong demand has kept prices high. Competitors like Super Talent have also begun to use the same NAND flash controllers from Indilinx, leading to shortages.

In response, OCZ has been looking for a second source for high-end NAND flash controllers. Samsung is a major supplier of NAND flash memory to OCZ, and has been seeking to broaden its product line by selling NAND controllers and DRAM cache as well. Samsung previously confirmed to DailyTech that they currently have no plans to sell their own PB22-J SSDs to the channel, preferring to sell them to OEMs instead. 

OCZ's new Summit series of SSDs use the same S3C29RBB01 controller as found in the Samsung PB22-J. All models will come with a 128MB DRAM cache, doubling that of the Vertex series. The 250GB and 120GB models have a maximum read speed of 220MB/s and a maximum write speeds of 200MB/s. However, sustained writes can achieve 200MB/s; double that of the Vertex series using MLC chips. The 60GB model has the same read speed, but only 125MB/s max write. Random read and write speeds were not available.

The company will compete against Intel in the enterprise market with the Summit series and the new Vertex EX series, which uses faster SLC memory. Retail availability of the Summit series is expected later this week. Pricing information has not yet been disclosed.

OCZ, however, is not the first partner to use Samsung's new controller. Corsair launched and shipped its P256 drive last week. It uses the same S3C29RBB01 NAND controller, but has different firmware.



Comments     Threshold


This article is over a month old, voting and posting comments is disabled

Internal "RAID"?
By descendency on 5/20/2009 10:42:06 AM , Rating: 1
Any info whether these drives will use internalized "RAID"?




RE: Internal "RAID"?
By Jansen (blog) on 5/20/2009 10:45:20 AM , Rating: 2
There is no internal RAID. The use of higher quality flash, multiple channels, and large cache negates the need for the setup that Apex drives use.


RE: Internal "RAID"?
By Ristogod on 5/20/2009 10:59:45 AM , Rating: 2
And the difference between multiple channels and RAID 0 are what?


RE: Internal "RAID"?
By Jansen (blog) on 5/20/2009 11:09:02 AM , Rating: 3
The older Samsung controller used 4 8-bit channels to access the flash, the new one uses 8 8-bit channels.

RAID 0 in a striping scenario would be use to connect the controllers together. It adds latency and overhead, and is not needed in this case.


RE: Internal "RAID"?
By Ristogod on 5/20/2009 4:08:25 PM , Rating: 2
Right.

My point was that fundamentally the concept is the same. Splitting data up in parallel. Whether the controller handles within, or an additional controller (RAID) handles multiple controllers itself, are simply different ways off implementing the concept.


RE: Internal "RAID"?
By Jansen (blog) on 5/20/2009 4:19:21 PM , Rating: 2
I respectfully disagree.

Adding a RAID controller introduces an additional point of failure, as does having the extra controller. It is an unnecessary additional layer.

It is much faster, easier, and cost effective to implement more channels on the controller than to have an internal RAID setup.


RE: Internal "RAID"?
By Alexstarfire on 5/20/2009 11:27:52 PM , Rating: 3
He's saying the concept is the same, not the implementation. Damn public education.


RE: Internal "RAID"?
By emboss on 5/21/2009 11:04:40 AM , Rating: 2
quote:
The older Samsung controller used 4 8-bit channels to access the flash, the new one uses 8 8-bit channels.


Any non-NDA'd refs for this? I've never managed to get any info out of Samsung regarding their controllers :(


RE: Internal "RAID"?
By therealnickdanger on 5/20/2009 10:54:32 AM , Rating: 2
Doubtful. I say that only because they are already hovering near the limits of SATA-3Gbps as it is (factoring for SATA overhead), internal RAID would only offer about a 10% improvement in speed while likely more than doubling the price. Internal RAID helps with slower SSDs with mediocre controllers, but the new Indilix and Samsung controllers work just fine without internal RAID.

You're much, much better off using true RAID. Until the industry goes mainstream with SATA-6Gbps or offers more PCIe solutions to meet the capabilities of SSDs, we're kinda stuck with a limits of 3Gbps per port.


RE: Internal "RAID"?
By descendency on 5/20/2009 1:18:53 PM , Rating: 1
My concern isn't max throughput but sustained throughput.

However as someone said, the latency makes it a less appealing idea.


RE: Internal "RAID"?
By winterspan on 5/22/2009 2:00:54 AM , Rating: 2
The high-end OCZ Vertex and now Summit series don't use RAID, but OCZ's older "Apex" model used an internal RAID0 setup with dual flash controllers to improve throughput because it used the crappy JMicron controller.

The Indilinx controller on the "Vertex" series and this new Samsung controller on the Summit, both with DRAM cache, do not really need RAID as the drives using them are the fastest (read+write combineD) MLC drives on the market. The only thing faster are the Intel drives, and only in read performance. The Intel X25M only does about 30-50% of the write speed of the Vertex and Summit drives.


"A politician stumbles over himself... Then they pick it out. They edit it. He runs the clip, and then he makes a funny face, and the whole audience has a Pavlovian response." -- Joe Scarborough on John Stewart over Jim Cramer














botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki