storage technology research

Google DRAM study turns conventional wisdom on its head…again

Remember the research paper Google made a splash with two years ago on disk drive failure rates? The one that showed that most failed drives didn’t raise significant SMART flags, failed to find a correlation between temperature and utilizaation with failure rates, and instead established that failure rates are more correlated to drive manufacturer, model and age?

Well, there’s now a DRAM equivalent — and it doesn’t paint a much prettier picture than the one on hard drive failures.

According to a new paper, “DRAM Errors in the Wild: A Large-Scale Field Study“, engineers from Google and the University of Toronto found that once again, failure rates and patterns did not match the received wisdom in the industry about how Dual Inline Memory Modules (DIMMs) behave. According to the paper:

We find that DRAM error behavior in the field differs in many key aspects from commonly held assumptions. For example, we observe DRAM error rates that are orders of magnitude higher than previously reported, with 25,000 to 70,000 errors per billion device hours per Mbit and more than 8% of DIMMs affected by errors per year. We provide strong evidence that memory errors are dominated by hard errors, rather than soft errors, which previous work suspects to be the dominant error mode. We find that temperature, known to strongly impact DIMM error rates in lab conditions, has a surprisingly small effect on error behavior in the field, when taking all other factors into account. Finally, unlike commonly feared, we don’t observe any indication that newer generations of DIMMs have worse error behavior.

As with the disk drive study, temperature also doesn’t play a huge role in DRAM failures. Here, vendor and model didn’t make as much difference as in the disk drive study.

However, the study showed errors were more highly dependent on motherboard design than previously thought. And contrary to conventional wisdom about DRAM, more failures were hardware than software-based. According to an article analyzing the paper by Data Mobility Group’s Robin Harris,

This means that some popular [motherboards] have poor EMI hygiene. Route a memory trace too close to noisy component or shirk on grounding layers and instant error problems…For all platforms they found that 20% of the machines with errors make up more than 90% of all observed errors on that platform. There be lemons out there!

These two reports raise one common question, according to Harris — why didn’t we know about these things before? As he put it, “Big system vendors have scads of data on disk drives, DRAM, network adapters, OS and filesystem based on mortality and tech support calls, but do they share this with the consuming public? Nothing to see here folks, just move along.”

IDC: Unstructured data will become the primary task for storage

According to a new IDC Enterprise Disk Storage Consumption Model report released this week, transaction-intensive applications are giving way as the main segment of enterprise data to an expanded range of apps as well as a tendency to create more copies of data and records for business analytics including data mining and e-Discovery.

The report estimates that unstructured data in traditional data centers will eclipse the growth of transaction-based data that until recently has been the bulk of enterprise data processing. While transactional data is still projected to grow at a compound annual growth rate of 21.8%, it’s far outpaced by a 61.7% CAGR predicted for unstructured data in traditional data centers.

“In the very near future, the management and organization of file-based information will become the primary task for many storage administrators in corporate datacenters,” the report reads. “And this shift will have a significant impact on how companies assess storage solutions in terms of systems’ performance, operational efficiency, and file services intelligence.”

The IDC report also builds on research first highlighted in an IDC blog last week concerning the cloud. According to the report, the sharpest growth in storage capacity will come from new organizations described as “content depots.” IDC estimates storage consumption from these organizations will grow at a compound annual growth rate of 91.8% through 2012. Examples of content depots  include the usual cloud suspects: Google, Amazon, Flickr, and YouTube.

These content depots have different IT requirements and infrastructures than traditional enterprise data centers. We’re seeing examples of these new  infrastructures pop up in the market, including systems with logical abstraction between the hardware and software elements; the use of commodity servers as a hardware basis for storage platforms; and the use of clustered file systems.

Some in the industry have compared this “serverization” of storage to the transition between proprietary workstations and PCs in the 1980’s. But IDC analyst Rick Villars says this isn’t a zero-sum game. “This isn’t going to replace traditional IT,” he said. “Ninety-five percent of what people are developing and building in the storage industry today is irrelevant to what the cloud is building. You could take that as a negative, but it also translates into opportunity. These are new market spaces and new storage consumers that weren’t around five years ago.”

There’s been a lot of discussion lately about the role the cloud will play as the global economy softens. There is a difference of opinion between those who see a capital-strapped storage market as an even more conservative and risk-averse one and those who argue the opportunity to avoid capital expenditures will nudge traditional IT applications into the cloud. Still others point out the hurdles to cloud computing that remain, including network scalability and bandwidth constraints.

For example, when it comes to storage applications such as archiving, analyst reports from Forrester Research this year cited  latency in accessing off-site archived messages and searching them for e-discovery as major barriers to adoption for archiving software-as-a-service (SaaS) offerings.

Cloud computing “definitely exposes weaknesses in networking,” Villars said, but “the closest point to the end user is the cloud, if you want to distribute content to end users spread around the world.”

Other challenges include the growing pains major cloud infrastructures such as Amazon’s S3 have experienced over the last 18 months, and the potential risk of putting more enterprise data eggs in one service provider’s cloud data center basket. Villars points out, “I doubt Amazon has had more problems than a typical large enteprise, and they offer backup with geographic distribution for free.”

However, geographic distribution brings with it its own challenges, such as varying regulations among different countries. “There are regulatory problems with Europe,” Villars said. “Laws there say that if you have data on a European customer, yhou can’t move it out of Europe. If you want your cloud provider to spread copies between the U.S., Asia and Europe for global redundancy, that becomes an issue.”

EMC shares vision of AI storage arrays

 
“Project Futon” demo area on the show floor at EMC World.
Complete with actual futon.

EMC has made a habit of opening the kimono lately, especially at this year’s EMC World where execs divulged detailed roadmap information around backup and archiving software consolidation.

They also had an exhibit set up as part of the show floor called the Innovation Showcase. The showcase displayed products on tap for much farther down the line, including Project Futon, a consumer appliance being developed at EMC’s R&D facilities in China to store digital photos.

Eventually, the Futon software would automatically upload photos to the “Futon Cloud,” according to EMC senior engineer Hongbin Yin. Yin was one of several engineers on hand to demonstrate their prototypes. The goal would be to make small home appliances a “local cache” for multimedia, with long-term storage taking place in the cloud. Futon is also being developed to automatically collect metadata, including the location photos that were taken, and to integrate that information into other applications like Google Maps.

Also on display was a diagram of “Centera with Data Lineage,” which, according to Burt Kaliski, senior director of EMC’s Innovation Network, would allow Centera to archive not just elements of a workflow but the workflow itself, and link documents to related data in spreadsheets and databases.

Senior consulting software engineer Sorin Faibish was showing off his own diagram of “Application-Aware Intelligent Storage.” This would combine artificial intelligence software capable of being “trained” with hardware-embedded VMware ESX servers to automatically spawn services like data migration, encryption and replication to data as it comes into the cache on a storage array. The embedded ESX host would run EMC’s RecoverPoint CDP inside, logging and catalogging I/O, indexing data for input into a modeling engine, which would then decide on the proper way to store and protect the data before flushing it to disk.

No time frame was given on any of the prototypes. Faibish’s project would require development of advanced artificial intelligence software using concepts like neural networks, fuzzy-logic modeling and genetic-based learning. “It could prevent commoditization of the storage array,” he said. “But it’s still a dream.”

Storage in high places

Two press releases caught my eye this week that aren’t exactly earth-shattering, but got me thinking about the way the storage market is changing and widening.

First, SanDisk revealed that its flash cards are recording footage of an excursion to Everest by a three-member climbing team sponsored by Dell, Windows Vista, MSN and MSNBC. Here’s a media gallery of the chilly-looking expedition so far.

Then there was also an announcement from RAID, Inc. of its compact Razor RAID array using 2.5-inch SAS drives, billed as “ideal for small spaces such as cockpits, tanks, submarines and other civilian applications with specific space constraints.” The ‘cockpits’ idea got my imagination going.

Between flash memory, with fewer moving parts and power requirements, and small-form-factor hard disks, not to mention the continued increase in content we store digitally, enterprise-level data storage is worming its way into unheard-of environments. As such, many in the industry have been predicting an increasing focus on edge devices, mobile computing environments and the mobile workforce for the storage market. Hopefully enterprise storage managers are paying attention to these new frontiers while architecting storage at headquarters.

Also, since it’s Friday, and who couldn’t use a laugh? Check out this priceless Gizmodo post on an internal Microsoft sales video that recently made its awkward YouTube debut. Key line: “You’ve gotta wonder how, in a company the size of Microsoft, there’s not a single person who [can] step up and say “Hey, you know what? This Vista music video we’re making for the sales department, complete with a cheesy Bruce Springsteen impersonator and horrible music, damages the dignity of not only everyone involved in its production, but everyone who watches it.”

The economy and technology innovation

I love listening to NPR. I listen to, watch and read many news sources, but I find the stories they choose and the nuances they bring to their reporting refreshing. I was listening to NPR this morning when a very rare thing happened–I heard someone being interviewed that I’ve interviewed before myself. It’s not often that IT industry news makes a mainstream general-purpose broadcast, so I paid close attention.

The pundit in question was Rob Enderle, a technology analyst I interviewed last month when EMC acquired Pi. After hearing his brief comments on the current state of the US economy and how he predicts it will affect technology innovation in Silicon Valley, I called him up myself and dug a little deeper into the matter with him.

Continued »

Interesting tidbits from around the storage blogosphere

The storage market is a vibrant one right now, and as social networking concepts like blogs become more popular in the corporate world, the storage industry has a lively, varied blogosphere to match. Below are examples of some of the more interesting commentary I’ve seen lately, in case you missed them.

Continued »

New SSDs are on their way

Or so says Pliant Technology, a new company that just received $8 million in Series A funding. It’s comprised of former execs from storage companies including Maxtor, Quantum, Fujitsu and Seagate.

The cast of characters is as follows:

• Jim McCoy, Chairman - Co-Founder of Maxtor and Quantum
• Amyl Ahola, CEO - Former CEO of TeraStor, vice president at Seagate and Control Data
• Mike Chenery, President/Founder - Former vice president of advanced product engineering at Fujitsu
• Doug Prins, Founder/Chief Architect - Former consultant for Fujitsu, Emulex, and Q-Logic
• Aaron Olbrich, Founder/CTO - Formerly at Fujitsu and IBM

And that’s just about all we know in detail right now about Pliant. I spoke with McCoy this week about the announcement of funding; he said the company has decided to come out of stealth now, but has been working on perfecting the solid-state drive for the last two years.

The new company is aiming to improve the solid-state drive with its products, which are due out by the end of this year, with alpha and beta testing scheduled beginning this summer. Pliant’s drives will perform better than current flash drives, “closer to what the DRAM people have,” McCoy claims. The drives have also been “designed for a 24×7 operating environment, with error rates equal to or better than hard drives.” Specifically, the drives are going to tackle an issue McCoy says has been a dirty secret in the solid-state game: read disturb, a phenomenon in which reading data from one portion of a flash drive causes degradation in nearby bits.

Existing solid-state vendors have tried to address this problem, as well as issues with write endurance, using error correction codes (ECCs). But according to McCoy, ECC is not enough. “ECCs are a minimal starting point,” he said. “By themselves, they are not sufficient.”

If that gets you all wound up about the state of solid state, though, you’re going to have to wait to find out how exactly Pliant plans to build a better mousetrap. The specifics of its technical approach are “confidential at this point,” said McCoy.

Will the new and improved Pliant drives be able to do anything about the acquisition costs that are keeping many users away from solid-state drives right now? “There won’t be much of a price penalty over other [SSD products],” McCoy said, which I’ll take as a no. McCoy did point out that long-term, solid state is more cost-effective than over-provisioning hard drives.

The problem is, users rarely start from scratch; many will have over-provisioned hard drives already, and would need to start by adding very expensive SSDs on top of already very expensive assets. “Customers are reaching the end of possible performance with hard drives,” McCoy countered. “And new systems [like EMC's Symmetrix] are going to start going out with a combination of drives.”

According to research from IDC, performance and mobility-related requirements will propel SSD revenues from $373 million in 2006 to $5.4 billion in 2011, a 71% CAGR. And I’ve heard many in the industry lament that while the capacity of spinning drives has been going up continually, the ability to get data off those drives faster is not keeping pace. Something will obviously need to change.

Meanwhile, the answer to the question of exactly how Pliant’s products propose to be a catalyst in that equation remains in stealth for now.

Open-source grid storage project goes commercial

Some of you may have heard of Cleversafe, until now an open-source research project working to develop the prototype of a system that would automatically spread data over geographically dispersed grids while encrypting it.

Cleversafe has been making slow but steady progress over the last year and a half or so and have been keeping me updated. Their concept, is an interesting one: a way to automate the “chunking” of data over geographically dispersed nodes through new algorithms that also make each chunk of data unreadable, essentially combining primary storage with disaster recovery and data security all in one go, as our friends across the pond would say.

So far, Cleversafe has launched itself as an open-source project, invited developers to play with the Dispersed Storage Network (DSNet) prototype, and signed up 14 internet service provider (ISP) partners to pilot the service. This spring, those partners will begin to sell some actual software and hardware to go with the pie-in-the-sky concept.

The new products, which will be generally available May 31, include a storage node, called the Cleversafe Slicestor; a storage router, called the Cleversafe Accesser; and a software management console called the Cleversafe Manager. Each Slicestor will hold 3 TB raw in a 1U pizza box. There is no formal restriction on the number of Slicestors and Accesser nodes in one grid, but the first products will be offered in groups of 8 and 16 nodes, with a 4:1 ratio of storage to router nodes recommended. The nodes can be kept in a single rack in one location or distributed globally. Cleversafe says its business model will be to offer its grids directly to enterprises, as well as ISPs and managed service providers who can offer Cleversafe storage as an online or hosted service.

This is the kind of stuff that really intrigues me in the storage market–the kind of stuff that makes me envision Conan O’Brien with a flashlight under his chin singing “In the Year 2000…” The futuristic stuff. As a general, all-around nerd, it’s interesting to me to talk to the people planning the next generation of technology, to learn what the challenges are and what goals their sights are set on. The Cleversafe concept is a particularly interesting one to me given the global-scale DR challenges we’re beginning to face.

When we chatted about it last week, though, Taneja Group founder Arun Taneja tempered my enthusiasm with the reminder that future products are just that: in the future, and the proof is in the pudding. “At the concept level I’ve never had any issue with Cleversafe,” he said. “But while the concept is interesting, provability will take a long time.” Cleversafe must show its product can support multi-tenancy environments reliably, without mixing up data chunks, and must show that its performance and ability to recover data are what it says they are.

And while some of the deepest innovations in technology are happening around storage, Taneja also reminded me that the market for storage products remains more conservative than most. “Even if Cleversafe can prove that this is the best thing since sliced bread, the GMs, Fords and Pepsis of the world would have to test something like this for years before they’d trust it,” he said.

So we might not be looking at The Storage Internet ™ anytime soon. But I’m going to keep watching.

Who’s going to pay for digital preservation?

I can hardly call myself a storage geek–I don’t know a MAC address from a Macintosh and couldn’t operate a CLI with a gun to my head. So it’s rare I take a personal interest, the way Tory does, in most of the products or trends I cover.

The one exception to that is the idea of digital preservation. This is probably because, unlike a true storage geek, I don’t have to worry about trying to fix a machine that’s broken or trying to throttle my service engineers. So I have time on my hands to think about the long-term future of data, data storage, and what we’re going to do with all the important records that are currently being converted from physical format to digital. Spinning disk still has nothing on a cave painting, data tapes have nothing on an acid-free paper book, and in 100 years, we might have an unprecedented historical problem: how to preserve our culture and our information for future generations.

That’s the kind of thing you don’t have to be able to architect a storage fabric to be affected by. Every living person has a vested interest in how the human race will pass on knowledge and information over the long haul.

The problem is, those of us with the time to think about this stuff aren’t the ones who know how to answer that question, and the ones with the know-how are too busy putting out day-to-day fires in their data centers to worry about how it’s all going to work when they’re long gone.

And who says it’s their (your) responsibility anyway? Shouldn’t institutions like the National Archives be the ones worrying about it? Shouldn’t the storage vendors be the ones developing the right media for long-term storage?

As of this week, there’s finally a publicly-funded consortium at least trying to find the answer to those questions about digital preservation, all leading up to the biggest mystery of them all: Who’s going to pay for it?

The consortium, known as the Blue Ribbon Task Force on Digital Preservation, was launched by the National Science Foundation and the Andrew W. Mellon Foundation in partnership with the Library of Congress, the Joint Information Systems Committee of the United Kingdom, the Council on Library and Information Resources, and the National Archives and Records Administration. The consortium, headed up by academics from the San Diego Supercomputing Center, will attempt to bring together testimony from a variety of sources — consumer and enterprise, vendor and end-user — to arrive at a sustainable economic model for digital preservation.

The group has been funded for a two-year project. The first year of the project, according to Francine Berman, Director, San Diego Supercomputer Center and High Performance Computing Endowed Chair, UC San Diego, and co-Chair of the Blue Ribbon Task Force, will produce a report on “a survey of what we know.” The initial report will feature case studies and opinions from experts in digital preservation, and is expected to appear by the end of 2008 or early 2009. By 2010, the task force hopes to have a second report suggesting an approach to digital preservation that’s the most cost-effective and logistically feasible for the most people.

It’s all a little loosey-goosey, Berman admitted, saying, “These are open questions.” So far the group doesn’t have much idea what its direction will be. Alternatives for economic models that will be taken into consideration include an iTunes-like pay-per-use model; a privatized model relying on corporations to finance preservation; or a public-goods model that preserves digital records the same way public parks are preserved, through a collective public trust.

Further complicating matters, “there won’t be a one-size-fits-all solution to the digital preservation question,” according to Berman. Consumers will be concerned with preserving family photos, for example, which will be an entirely different process from preserving corporate and government records. Preserving digitally-recorded works of art and multimedia files will be yet another issue to resolve.

Personally, I’m a little reluctant to put much stock in a government study until I see it produce actionable results, and as a taxpayer I’m not nuts about the number of studies my hard-earned dollars go to that just tell us things we already know. But in this case, I’m just happy someone’s thinking about it. And maybe getting others to start thinking about it a little more, too.

Raising awareness is another goal for the task force, Berman confirmed. “My dry cleaner knows what global warming is, and could also probably give you a basic definition of the human genome,” she said. “What we’re looking for is that same level of understanding about digital preservation, which also affects us all.”

Buffalo unleashes 100 GB Flash drive

Even my friends who don’t normally follow the storage business are atwitter over an Engadget report that Buffalo has unleashed a 100 GB behemoth flash drive upon the world. Geeks everywhere are probably salivating to take the thing apart (yes, I’m looking at you, Tory) … unfortunately, they’ll have to wait. The catch is that Buffalo is only releasing the product for now in its home country of Japan.

According to company reps, the $1,000 asking price for the credit-card sized USB accessory makes it less than cost-effective to import right now. (If you just can’t get enough flash memory, there are 64 GB monsters roaming North America.)

The Engadget comments section also contains an interesting discussion of the merits of such a large flash drive. In the Engadget screenshot, the card looks like a  behemoth, but the post says it’s about the size of a credit card. Still, it launched a spirited discussion that I think asks some pertinent questions, namely, “would it not be more practical to just buy a $300 travel drive?”

At this juncture, and at this price point, certainly. But Moore’s law waits for no man, and the price of a 100 GB card will come down. Hence the other questions that this announcement begs: at what capacity and price point does a mechanical drive become more practical than a solid state drive? How will that equation change over time? It’s something we in the storage market are going to have to examine more closely in the coming year.