disk drives

New standards emerge for power consumption testing and SAS connectivity

In a week chock full of product news from Storage Networking World (SNW) and elsewhere, some new standards have slipped in under the radar that may become important once the dust settles.

The first of these is the announcement of a new Storage Performance Council (SPC) benchmark for testing the power consumption of storage devices in the data center. The new SPC-1E spec follows the SPC-1 C/E spec announced in June. Where the SPC-1C/E spec covered storage components and small subsystems (limited to a maximum of 48 storage devices in no larger than a 4U enclosure profile), the SPC-1E spec expands that support to include larger, more complex storage configurations.

According to an SPC presentation on the new benchmark,  ”SPC-1/E is applicable to any SPC-1 storage
configuration that can be measured with a single SPC approved power meter/analyzer.”

For more on how the SPC-1C/E and SPC-1E benchmarks work, see our story on the SPC-1C/E announcement. Users should especially be aware of the parts of the benchmark calculation that can only be specified by vendors.

Still, even an approximate or idealized lab result for power consumption of storage systems would be an improvement over the tools avialable to reliably spec power consumption, increasingly a key cost factor for data centers that users in economically strapped times are looking to cut.

***

Speaking of cutting costs, Serial Attached SCSI (SAS) devices are widely regarded as the cheaper choice of the future to replace Fibre Channel systems. With 6 Gbps SAS products now beginning to ship, the SCSI Trade Association laid out its roadmap for the future of connectivity between Serial Attached SCSI drives and other elements of the infrastructure.

3 Gbps SAS devices connected via InfiniBand connectors; the Mini-SAS HD connector will be used with most 6 Gbps devices. The new roadmap laid out this week specifies that the Mini-SAS HD connector will be the hardware of choice going forward for all types of connectivity into SAS devices.

Why do you care? Because the development plans for the Mini-SAS HD connector going forward will allow it to serve optical, active and passive copper cables with one connector device, and automatically detect the type of cable it’s attached to — meaning that by the time 12 Gbps SAS rolls around, less hardware wil need to be ripped and replaced to support it. Another thing the connector will support in the future is managed connections, meaning a tiny bit of memory in the connector itself that allows the devices to be queried for reporting and monitoring.

The ability to connect SAS devices over optical and active copper cables is a pretty big deal — cable length and expandability limitations have improved significantly with SAS-2, but native cable lengths currently remain limited to 10 meters. While this is already making data center SAS subsystems a reality, it will need more robust connectivity attributes to compete directly with Fibre Channel. Optical cables can stretch as far as 100 meters, and active copper (so called because it contains transcievers that boost signals) to 20 meters.

Alliance pumps new life into Plasmon

Alliance Storage Technologies Inc. (ASTI), the Colorado company that bought Plasmon’s assets last January, says it is selling the Plasmon product line whole and hopes to significantly expand its business based on the proprietary Ultra Density Optical (UDO) technology.

ASTI was a Plasmon reseller before the U.K.-based archiving vendor went under after years of financial problems. ASTI picked up Plasmon’s assets for an undisclosed sum, leased Plasmon’s Colorado Springs manufacturing plant, hired many of its employees and is now stepping up marketing of Plasmon products. ASTI will keep the Plasmon brand name and is selling its UDO appliances, drives, libraries and media after rebuilding its channel with new VARs and integrators.

“It’s an identical product lineup as Plasmon’s,” said Bill Gallagher, a former Plasmon exec who is now ASTI’s director of strategic accounts and regional sales director. “I don’t think Plasmon’s failure was a failure of technology. The company suffered for years with restructuring and trying to get its financials in order. Alliance is profitable, and we haven’t seen any change in demand. Customers are happy, they wanted to see what would happen. ”

ASTI has sold optical storage for more than 10 years, carrying products from Hewlett-Packard, IBM, and Sony as well as Plasmon. ASTI CEO Chris Carr says the company is committed to the future of UDO. “Last year we caught wind of Plasmon’s financial difficulties and we saw an opportunity,” he said. “Specifically, we were looking for UDO technology.”

UDO discs hold up to 60 GB and are supposed to last for more than 50 years. Plasmon’s largest libraries have 638 slots and store 38.3 TB. ASTI execs claim Plasmon shipped over 17,000 libraries. ASTI will not honor service contracts for Plasmon customers but is offering discounts on new contracts, Carr said.

Fusion-io builds SSD bridge between SLC, MLC

Fusion-io has taken a step towards bridging the gap between expensive single-level cell (SLC) and cheaper but slower and less reliable multi-level cell (MLC) NAND Flash.

The startup calls the new solid state drive (SSD) technology single mode level cell (SMLC) and expects to be shipping products in its ioDrive and ioDrive Duo PCI Express product lines this quarter.

Fusion-io says SMLC “combines a cost-effective MLC-based solid-state solution with the endurance and performance of SLC,” but it’s really a third option that falls between SLC and MLC in price and performance.

Fusion-io hasn’t released performance numbers, but CTO David Flynn says the SMLC drives close the gap in write speeds and endurance cycles between MLC and SLC. SMLC drives store two bits per cell and come in capacities of 160 GB and 320 GB just like MLC drives – although the SMLC drives require greater overprovisioning to reach those capacities. Generally, SLC drives write about 20% to 30% faster than MLC drives and have about 10 times the write cycles. For the most part, MLC’s shortcomings have kept it out of enterprise SSD products while SLC’s price still scares off a lot of people.

“(SMLC) is very close to SLC,” Flynn said. “I wouldn’t say it’s exactly SLC, but it’s sufficiently close for most uses cases.”

Flynn says SMLC drives will roughly split the difference in cost between its performance SLC ($30 per GB) and capacity MLC ($15 per GB) drives.

Fusion-io already ships enterprise MLC drives that Hewlett-Packard sells as the HP StorageWorks 320GB IO Accelerator.

“This (SMLC) is subtly different,” Flynn says. “Now we can get endurance and performance characteristics of SLC.”

The difference is in the way the controller manages the NAND Flash, he says. “We don’t need special MLC Flash, that would defeat the purpose,” Flynn said. “The purpose is not to have special requirements.”

Dell also sells Fusion-io cards, and IBM has released test results and is committed to selling Fusion-io SSDs down the road.

None of Fusion-io’s partners have publicly signed on to the SMLC cards yet, but Enterprise Strategy Group analyst Mark Peters says SMLC will likely become a third category for NAND Flash alongside SLC and MLC, at least until NAND is replaced by better technology.

“More people will follow, because they have to,” Peters says. “It’s logical. Every piece of research we’ve done says the No. 1 reason people aren’t adopting solid state is price, and this is a move to get the price down.”

Flynn agrees that Fusion-io won’t be the only vendor with SMLC, even if others call it something different.

“We’re first, but we don’t think we’ll be the last,” he said. “It’s too compelling.”

Samsung adds encryption to consumer SSDs; Dell to ship in notebooks

Samsung is claiming it’s the first to ship a consumer solid state drive (SSD) with full-disk encryption (FDE) through a new partnership with security vendor Wave Systems Corp. The 256GB, 128GB, and 64GB SSDs will be available in both 1.8-inch and 2.5-inch form factors. Dell says it will ship the drives in its Latitude line of desktops and notebooks.

Samsung’s drives generate and store encryption keys and access credentials are in the drive hardware, and they are never held in the operating system or by application software. When ordered in a new computer, the drives will come bundled with Wave’s Embassy Trusted Drive Manager software for life cycle management of the drive. Teh software includes pre-boot authentication, enrolls drive administrators and users, and enables backup of drive credentials. Available separately, Wave’s Embassy Remote Administration Server allows an IT administrator to remotely turn on SSDs and adds event logs for compliance.

It probably won’t be long before full-disk encryption also hits the enterprise SSD space. It’s already working its way in on the spinning-disk side, where it’s being pushed by drive maker Seagate, controller maker LSI and systems vendor IBM. Multiple converging standards for key management are also being developed for the enterprise.

WD dives in to SSDs

Drive maker Western Digital is looking to hop on to the solid-state drive bandwagon by paying $65 million for a company that has the technology.

WD today completed a cash acquisition of SiliconSystems, Inc. of Aliso Viejo, Calif., a supplier of solid-state drives for the embedded systems market. The SiliconSystems’ product portfolio includes solid-state drives with SATA, EIDE, PC Card, USB and other interfaces in 2.5-inch, 1.8-inch, and other form factors. WD plans to develop new solid-state offerings to be embedded in OEM systems with the new intellectual property.

Depending on who you talk to, SSDs could become the new tier 1 in storage over the next several years. Not everyone shares such a bullish outlook, but it’s clear to drive makers like WD and Seagate that they must have an offering in this space to compete. Forward Insights president and analyst Gregory Wong sees the acquisition as the only way for WD to develop a competitive offering at this point, noting on his blog today that WD “dabbled” in SSD technology in the early 1990’s.

“However, it appears that early experience wasn’t enough for a latecomer to catch up to the vast improvements in performance and reliability made by SSD vendors recently,” he wrote. “System-level solutions particulary on the firmware side are required to manage the increasing complexities of NAND flash with each new technology generation.”

Wall Street analyst Aaron Rakers of Stifel Nicolaus Equity Research wrote in a note to investors Monday that he expects to see the combined entities focus on positioning embedded SSDs in blade servers.

Dell/EqualLogic prepares SSD IP SAN

Dell is about to add solid state drive (SSD) support to its EqualLogic iSCSI SANs in a new PS6000 model. As first reported on ChannelWeb last Friday, the PS6000 will also have four Ethernet ports, one more than EqualLogic PS5000 arrays have.

Dell officials did not return requests for comment today by Storage Soup, but several industry sources have confirmed the report is accurate and say the PS6000S will support 16 solid-state drives. A PS6000E will also be available with only SATA drives, according to one customer who asked not to be named because the product has not yet been formally released.

The general opinion on solid state is that customers will hold out for higher capacities and other features before they buy. Several EqualLogic customers reached by Storage Soup today said they still found SATA drives adequate for their needs.

However, according to Alan J. Hunt, Manager of Operations for Dickinson Wright PLLC, “It’s just the beginning of the market. In a few years I suspect [SSDs are] all we’re going to have–it’s kind of the beginning of the next big wave.”

Hunt added that a fourth port on EqualLogic’s arrays could be more significant than it might appear. “A fourth port means you would have balance if you have two switches and want redundancy,” he said. “Or you could make it a dedicated management port and still have three ports.”

Missing from the coming product update, if reports are accurate, would be 10-Gigabit Ethernet support, which a Dell spokesperson said last year is on the EqualLogic roadmap for 2009. But like with SSDs, EqualLogic customers and resellers say 10-Gig Ethernet can wait.

“I haven’t seen a susstantial interest in 10 Gig,” said Broadleaf Services account executive and EqualLogic VAR Christopher Baer. “There aren’t a lot of applications that require that kind of throughput yet.”

Other customers say 10-GigE would future-proof the array, even if they don’t need the bandwidth quite yet. “Why add another port? Why not 10-GigE and really get this thing going?” said a customer in the education field who requested that his name not be used as he is not authorized to speak with the press. Some pieces of the IT infrastructure in this user’s shop have been upgraded to 10-GigE already, including the network backbone.

“I don’t personally need the bandwidth,” Hunt said. “But it could add the ability to reduce the number of cables and passthrough modules for blades, as well as greatly simplifying VMware deployments.”

Riverbed contemplates SSDs, dedupe

Eric Wolford with his lobster bib.

Riverbed held a roundtable discussion about the company (and some other topics) with journalists last night at Boston’s Oceanaire restaurant in Government Center. I sat next to Eric Wolford, Riverbed’s SVP of marketing and business development, and he opened a conversation by asking me “What’s hot?”

“Well,” I told him, “these days I sometimes feel like I’m writing for SearchSSD.com rather than SearchStorage.com.” I didn’t expect Riverbed to be getting into the solid-state disk game, but Wolford said there’s probably a place for SSDs in at least some of its WAN optimization products, too.

Wolford said some large Riverbed customers use Steelhead devices on both sides of the wire for replication. At extremely high bandwidth (OC12 and above), Wolford said SSDs could help keep up when large volumes of data hit the devices’ disks simultaneously.

“With large data center-to-data center replication, they sometimes need so many spindles there’s an opportunity for solid-state storage,” he said.

But he doesn’t see SSDs replacing spinning disk systems, for Riverbed or the industry at large. “It’ll give us a new high end,” he said. And, he added, “an enormous amount of our business is at the T1 level and there’s really no opportunity for it there.”

Wolford also gave me an update on the Atlas primary storage dedupe product Riverbed was originally going to ship this year but recently pushed out until 2010.

“We got critical feedback from alpha customers where they want to deploy [Atlas], but don’t want dependency on the Steelhead appliance.” Wolford said. So Riverbed is working on bundling the Steelhead functionality into the Atlas product itself.

Atlas will sit out of band, he said, “to the side” of the array and perform post-process dedupe. Wolford says customers are hot for primary storage data reduction, but most vendors still can’t deliver it at speeds fast enough for primary storage. “If the device is out of the path of hot data, the performance burden isn’t as extensive,” he said.

A technology deep-dive with Fusion-io

We interviewed Fusion-io Inc. CTO David Flynn for one of our news stories today–here’s some nitty-gritty bonus footage on how the company’s product goes about protecting data, and how that compares to spinning-disk systems.

Beth: So one ioDrive is 320 GB. Is data striped across all the chips or do you have separate data sets?
Flynn: Each one of the Flash modules looks like a volume and you can either stripe them or mirror them to make them look like one volume. Or is you have multiple cards you can aggregate all of those volumes with RAID 10. We have RAID-5 like redundancy on the chips, then RAID between the memory modules. What we’ve come to realize after we introduced FlashBack is that it actually lets you get more capacity.

Most SSDs are 64 GB at most—32 GB, 64 GB. With this technology we put five to 10 times as many chips within our card. That would increase the failure rate because the individual chip’s failure rates add up. With our ability to compensate, we can get to higher capacities, and with that we can increase endurance, because you can spread the data out.

Internally it’s more like RAID 50 because I have eight die in my redundancy chip. There’s one parity die for each package. It’s 24+1 and then that quantity times eight, because there’s eight of those sets. If you were to line it up like disk drives, it would look exactly like that, 24 disk drives and then an extra one, 8 rows. So when we talk about this as a SAN in the palm of your hand we really mean it, because we’ve taken die within the various NAND packages and arrayed them together just like a disk array. It’s also self-healing in that if you have a fault the system reconstructs the data that otherwise might’ve gone missing and moves it to a different spot and turns off the use of the spot that failed. You don’t have to service it. It automatically just maps it out. Like Xiotech’s ISE product—that’s bleeding edge stuff for disk arrays, and it’s built into the silicon here.

What about double parity protection? That’s all the rage in the disk drive world these days. What if more than one die fails at once?
For us to rebuild and heal takes a split second. Having a second failure during that time is not going to happen. It takes so long to rebuild a disk drive—it can take more than a day now—that the probability of a double failure goes up. The other thing is that disk drive failures are often highly correlated—the drives come from the same batch. They tend to fail randomly but close to each other in time. Our portfolio does cover n+m redundancy as well as N+1 because we anticipate a day when we’re putting not hundreds of these die on the boards but thousands and going into the tens and hundreds of thousands.

At the same time the Flash memory has finite write endurance, so they are all going to wear out at some point. So how do you compensate for that?
We account for how many write cycles it’s been through so we can give somebody a running…like an odometer, for tread wear on a tire. You can go five years or 50,000 miles. We warranty it, and you can swap out the modules without needing a new carrier card. Because we have such high capacity we naturally get a longer lifespan. It’ll last for 5 years even if you’re doing nothing but writing constantly. Wear-out has been overrated I think because most of the failures people are seeing have nothing to do with wear-out, they have to do with internal events that cause chips to lose data.

Like what?
Here’s the four factors. This is the dirty little secret of the NAND world—it’s the newest fab process, which means it has its kinks. It’s the tightest feature size—they’re going to 32 nm. The density of the array of cells is achieved by sharing control lines. And then, fourth, and the real killer, to move the electrons into the floating gate cell it takes 20 volts internally. Most core voltages are well under a volt nowadays.

These four factors mean having a short-out event on one of these tiny little control lines—if you have just one chip it’s no big deal, it’s 40 out of a million. Which for a thumb drive, nobody would notice—it’s more likely to get shorted out in your pocket. But when you put hundreds of them together, now you have hundreds of those 40 out of a million chances to have something go bad, and that actually adds up to be something like one or two percent of these things fielded would have a data loss event. For a normal SSD the way they compensate is to put fewer chips on it or try to sweep the problem under the carpet—what they say if you talk to them is, ‘Well, we screen it very well, we run it in advance to make sure it’s not going to happen.’ You can screen it up front but there’s still probabilities of failure.

Here’s the thing: disk drives wear out, too. The trouble is, it’s unpredictable. One of the strongest motivators to going to solid state technology is the predictability of when you’re going to need to service it. And after a couple of years, you’re going to be able to replace it for a fraction of what it cost initially.

Fujitsu transfers hard drive, SSD business to Toshiba

Toshiba Corp. and Fujitsu Limited today made official what has been long rumored — Toshiba will take over Fujitsu’s hard disk and solid state drive business.

Toshiba will first take an 80% stake in a new company to be created by Fujitsu, while Fujitsu will maintain a 20% stake. The business will become a wholly owned subsidiary of Toshiba. The deal is expected to close this quarter, according to Scott Maccabe, vice president and general manager of Toshiba’s storage device division as well as a former senior manager with Fujitsu.

Maccabe said the down economy had Fujitsu looking to cut out non-profitable businesses. It has sold off its media businesses, including R&D and global sales organizations, to Toshiba, and has plans to sell its drive head business to Showa Denko K.K. (SDK). Fujitsu will now be focused on its systems and services businesses in storage, Maccabe said.

According to a Toshiba press release:

The consolidation of the two companies’ HDD businesses will enable Toshiba to reinforce its already strong position as a leading vendor of small form factor HDDs…widely used in notebook PCs, mobile devices, automotive and consumer electronics. It will also give Toshiba entry into the enterprise HDD market for server and data storage system applications, where Fujitsu is currently a leader…entry into the enterprise business will allow Toshiba to further enlarge its market-leading solid state drive (SSD) business by developing SSD products for servers and enterprise storage systems, fusing Toshiba’s NAND flash memory technology with Fujitsu’s enterprise HDD technology.

According to a note sent to investors by Jayson Noland of Robert W. Baird & Co. today, “The acquisition reduces mobile HDD suppliers to five from six, which should reduce industry capacity and improve pricing behavior … though we do not expect competitive dynamics in [the] enterprise to meaningfully change as a result of the merger.”

Apple founder Wozniak joins solid-state drive company

SSD maker Fusion-io announced today that Apple founder Steve Wozniak has joined it as Chief Scientist. According to a Fusion-io press release, “Wozniak will act as a key technical advisor to the Fusion-io research and development group. He will also work closely with the executive team of Fusion-io in formulating a company strategy that will accelerate the expansion of major global accounts.”

Wozniak is a big name for an emerging company like Fusion-io to land. Sometimes known as Apple’s “other Steve,” Wozniak is credited with significant engineering contributions to the personal computer revolution of the 1970’s.

Interestingly, Wozniak is not the only former Apple exec who has found his way to the storage–and specifically, solid-state storage–industry. Michael Cornwell, now heading up NAND business development for Sun Microsystems, was previously the manager of storage engineering for the iPod division of Apple.

Fusion-io came out of stealth last March with a PCIe flash card designed to give off-the-shelf servers SAN-like performance. Fusion-io calls its product the ioDrive, and it’s NAND-based storage that comes in 80 Gbyte, 160 Gbyte and 320 Gbyte configurations. The ioDrive fits in a standard PCI express slot, shows up to an operating system as traditional storage and can be enabled as virtual swap space. IBM announced last fall that it will be partnering with Fusion-io to add ioDrives to its servers, which may include the servers that run its SAN Volume Controller (SVC) network-based storage virtualization product.