Paul goes on to quote me, saying I “didn’t think primary data de-duplication technology was ready for prime time just yet.” I absolutely did say these words, but I am not sure if my point came across.

I’ve recently expounded about the benefits of deduplication technology, but have warned that it might not be all it’s cracked up to be in primary storage environments. By “primary” I mean those storage environments serving mission-critical applications. Although dedupe works great for backup and archiving, the random I/O, low latency, and high throughput of primary storage (and especially virtualized servers) might be too much for current systems. And as of now, only NetApp, Riverbed (soon), and startups greenBytes (see my story) and Ocarina (more on them another time) were willing to go on record with me as supporting deduplication of primary storage.

So what I meant was that deduplication is not yet ready for prime time in primary storage applications. No one should hesitate to use the technology for backup or archiving at this point, but make sure you do a thorough evaluation of the specific product you are selecting to make sure it delivers the performance you require!

© sfoskett for Stephen Foskett, Pack Rat, 2008. | Is Deduplication Ready for Prime Time?
This post was categorized as Enterprise storage, Virtual Storage. Each of my categories has its own feed if you'd like to filter out or focus on posts like this.

Although deduplication of storage is nothing new, with Data Domain and other making hay with the technique for years, it has never been ready for prime time – reduction of active primary storage applications like email and databases. Instead, deduplication has been relegated to second- or third-tier status, deduplicating archives and backup data. But change is in the air, and deduplication vendors are starting to bustle towards the bright lights of primary storage.

Stone Knives and Bear Skins

We have all been here before, of course. Back at the dawn of the personal computer era, data compression was a hot topic of conversation. I recall being so impressed by an article in Byte (1986:5, p99) outlining Huffman coding that I tried cooking up an implementation in Atari BASIC. Lossless compression has a magical pull to the geek in many of us – redundant data just wants to be eliminated!

Stacker dominated the disk compression world - until Microsoft introduced DOS 6.0

Companies soon applied compression to primary storage, especially the limited storage in personal computers. Stacker was a hit after 1990, until Microsoft built a workalike, called DoubleSpace, into DOS 6.0 in 1993, leading to a historical lawsuit. I personally used the ADDSTOR disk compression built into DR-DOS 6.0 to stretch two more years out of the 20 MB MFM hard drive in my AT&T PC6300 at WPI.

But something funny happened in the late 1990s: Compression began to lose its luster. Compressing data always takes quite a bit of CPU power, but this was offset somewhat by the truncated data transfers and more-efficient file system layout afforded in early PCs. But as disks got larger and faster, using precious CPU time to save space seemed less and less compelling. Today, although nearly every operating system includes built-in compression of files, folders, or perhaps disks, these features are rarely used. And compression was never popular in the performance-sensitive enterprise space.

Deduplication Has a Nice Ring

Although traditional fine-grained compression has not been very successful in the enterprise, its lanky cousin, single-instance storage, has long found niche jobs. Applications from databases to email systems to file servers have long had the ability to recognize to requests to store the exact same file or record, and to store just a single instance in this case. Even file systems have the ability to do single instance storage through the use of links, though this is initiated by the user rather than in an automated fashion.

In the late 1990s, FilePool began developing a content-addressable storage device, which was acquired by EMC in 2001. This device, later known as the Centera, was one of a number of storage platforms targeted at the archiving market introduced this decade. At the same time, virtual tape libraries made the jump from the mainframe to open systems. Both devices, being outside the critical path of performance but offering massive capacity, were well-suited to implement advanced capacity optimization technologies that combined the concepts of compression with single-instance storage. Thus was created the modern world of data deduplication.

What we think of as deduplication is neither fish nor fowl: It assesses larger “chunks” of data than compression technologies, delivering greater capacity savings and potentially reducing performance impact, but is more flexible than single-instancing, recognizing the similarities within files or objects.

But it is still maddeningly difficult to scale deduplication while maintaining performance. Rather than fight to maintain reasonable write throughput, most deduplication products have switched to post-processing, deferring their work to quieter times.

It’s Not Just for Breakfast

Regardless of their methods or underlying technology, no deduplication vendor has stood up to support challenging low-latency or high-throughput production applications, however. NetApp was the first to raise the issue of support for production applications, but although they tout the technology for VMware, they haven’t exactly been shouting from the rooftops to get their A-SIS deduplication technology deployed in other high-I/O applications. And I haven’t seen Hifn’s card yet.

Yesterday, I mentioned that greenBytes was adding deduplication to their ZFS-based storage array for primary data. And now Riverbed has fired another shot over the bow, repurposing their (deduplicating) WAN accelerator product for primary (file) storage. They might be able to pull it off, too, since they have a long list of customers who are already enjoying the technology in production. It’s not a stretch to suggest that Riverbed’s appliances can scale to handle production data loads. Although it’s file-only, I can imagine quite a few scenarios where this tech could really yield benefits. Could we come full-circle, with deduplication finally reaching the enterprise storage world?

© sfoskett for Stephen Foskett, Pack Rat, 2008. | Deduplication Coming to Primary Storage
This post was categorized as Computer History, Enterprise storage, Features, Virtual Storage. Each of my categories has its own feed if you'd like to filter out or focus on posts like this.

Spun Down

Although ZFS’ universal storage pool with non-RAID is a great concept, it stands in the way of at least one (sometimes) desirable storage technique: disk spin-down. Put simply, since every disk contains metadata, all disks must always be spinning. This issue is by no means a ZFS-only problem, though – certain vendors tout the (laughable) greenness of their storage systems, while hoping that the average user won’t notice the truth: That a disk simply cannot spin down while any part of it is in use. This means that tacking spin-down onto a regular storage array is like painting it a different color: There is no benefit whatsoever to the average user. Sure, a few non-provisioned drives might spin down, but what are you doing buying a lot of non-provisioned drives anyway?

The solution has always been right in front of everyone: Develop a new type of non-RAID with enough intelligence to allow drives to spin down when not used. This is what COPAN Systems did with their MAID technology: Invent an entirely new storage array, with integrated data protection and management techniques that allow alive but not active drives to spin down. Spin-down is not MAID any more than a bicycle is a Ducati.

Let’s make one thing clear: It’s really hard to reduce the power demands of storage devices. Disks guzzle watts like few other data center devices, and enterprise storage uses lots of disks. Lots of vendors are looking to hop onto the green storage bandwagon, and they all seem to realize that bringing some intelligence to power management by enabling spin-down is an open door. But it’s awfully hard to maintain performance and data protection when disks are spinning up and down all the time.

One element of the greenByte story is the way in which they have tweaked ZFS to allow disks to spin down. They limit the metadata updates to just a few disks, so the others can be idled when no access to them is made. The company suggests scheduling this for off hours to minimize latency as drives are brought back online, an approach that is less than optimal from an energy perspective but demonstrates that they understand just how difficult this problem is to crack. The core is there, however: They have integrated the data protection and storage management elements to enable spin-down to be practical.

Compressed

Another major storage industry theme of the last few years is deduplication of data. An advanced (or devolved, depending on your perspective) form of compression, deduplication allows a storage array to store duplicate data more efficiently, reducing the amount of capacity required for some applications. Data Domain is top-of-mind in this space, but just about everyone now offers some form of deduplication technology.

One major roadblock on the way to deduplication (or compression) nirvana is performance. Simply put, it’s really really hard to process data on the fly without affecting performance, especially as data scales up to the multi-terabyte range or as systems scale out to include multiple devices. One approach to tackling this issue is post-processing dedupe, which accepts incoming data in the normal way but goes back and processes it later to remove duplicates. This is the method NetApp uses, and they have leveraged it to become the first vendor to support deduplication of production applications.

Predictably, deduplication is another technology integrated into greenBytes’ “ZFS+” technology. They claim that they can handle inline compression at wire speed, and also claim deduplication inline. It’s not yet clear exactly what the difference between compression and deduplication is to the company, or just what kind of performance their inline technology will yield, but it’s certainly nice to see this tech integrated with ZFS!

Thin is In (the House!)

greenBytes gets closer to enterprise storage bingo by adding thin provisioning to the mix. Actually, as the company’s CTO was quick to point out, they had to offer virtual or thin provisioning to enable the rest of the system to function. When your storage is sliced and diced by their Cypress array, the only way to present storage is with a wink and a promise of capacity to spare. Thankfully this is not the core of their pitch, however.

The company also promises snapshots and CDP replication, all leveraging ZFS at the core. All they need to add is tier-0 solid state storage to get five chips in a row without even using the free space! Although greenBytes is using Sun’s Thumper chassis currently for their Cypress array, their core technology is the ZFS+ software, and I expect we might see this mixed quite differently in the future. This is a software company, not an array vendor.

All considered, greenBytes has thoroughly broken the link between physical and logical storage, and I applaud them for it. This is exactly the kind of storage revolution the industry needs right now.

© sfoskett for Stephen Foskett, Pack Rat, 2008. | greenBytes Embraces and Extends ZFS
This post was categorized as Enterprise storage, Virtual Storage. Each of my categories has its own feed if you'd like to filter out or focus on posts like this.