In-Depth

Still Another Data Warehouse Appliance Is Coming!

• By Stephen Swoyer
• 05/23/2007

If Paul Revere were alive and cruising the BI trade show circuit, he’d probably put it like this: Still another data warehouse (DW) appliance is coming! Still another data warehouse appliance is coming!

That was the big news out of last week’s TDWI World Conference in Boston, where DW appliance player Dataupia Corp.—an upstart company with anything but an upstart management team—announced its long-awaited entry in the appliance wars. Like its DW appliance kith, Dataupia’s Satori Server 12000 boasts integrated server, storage, and DW software. Unlike those other offerings, however, Satori Server purports to be a 2.0 appliance—that is, says Dataupia CEO Foster Hinshaw, Satori Server can plug right into an existing data center environment as-is. As far as relational databases and other data sources are concerned, he says, it’s business as usual.

"As the guy who came up with the moniker ‘database appliance,’ it has to be able to pass what I call the ‘Tivo test’—it has to be an integrated unit. You don’t figure out in a toaster what coil you use, what the plug looks like, whether you put a fuse on it or not. You just plug it in and use it. It just works," says Hinshaw, who—for the record — first coined the term "database appliance" during his stint with Netezza Inc., where he was co-founder and CTO.

Hinshaw’s "Tivo Test" is one of two differentiators which he says help distinguish second generation (or "version 2.0") appliances like Satori Server from their predecessor devices. The second, Hinshaw maintains, is improved scalability and performance: "In the second generation, we have more than enough capacity and power to handle any workload that’s out there."

That’s a heady claim, mostly because it serves to place Dataupia in somewhat rarified company: only Teradata (a division of NCR Corp.), IBM Corp., Oracle Corp., Sybase Corp., and a handful of other vendors can credibly make such a claim. It’s a claim Dataupia vice-president of marketing (and, like Hinshaw, Netezza veteran) Samantha Stone qualifies to some extent by noting that Dataupia envisions an enterprise in which all users—from hotshot-power-user-all-stars down to task-based workers—are exposed to DW features or services. This is one sense, she says, in which Dataupia is thinking Big Time.

"As appliances have matured, the value of them should be extended to more people in the organization, across more purposes than just the complex analytical environments," Stone argues. "There’s a value and a benefit to [exposing a DW to] a much larger number of people in an organization."

That’s one consequence of the DW appliance having effectively won the battle for industry recognition, Stone indicates. "The debate used to be ‘What is an appliance? Does it have any viability?’ For the most part people now understand that the appliance has a valuable contribution to make. They’re also discovering that they can deploy [appliances] across more places in an organization to complement their infrastructure," she says.

Hinshaw says the DW appliance can alter the manner in which organizations do data warehousing. "There’s been decades at this point of conversations about what an enterprise data warehouse is. Fact is, most people have implemented independent data marts. That’s because the way budgeting is handled in the different lines-of-business, there’s really no incentive for them to push for that [single enterprise DW]," he explains. "But if you can create an infrastructure to meet what that vision is—if you can show them how this helps execute on this vision—then that changes the discussion. If you can show them that they can still do their particular [line-of-business] activity against that [enterprise data warehouse] without having to change what they’re doing, so many more people in an organization are going to be able to benefit from that, then they become a lot more receptive."

Dataupia’s Satori Server, like competitive offerings from Netezza and DATAllegro Corp., is based on commodity hardware. (Netezza’s is a special case in this regard. It uses PowerPC-based processing engines, which aren’t quite as commoditized as DATAllegro’s and Dataupia’s Intel-based processing innards.) Like its appliance kith, Satori Server is actually a collection of blades, each of which is outfitted with 2 TB of storage. Customers add additional capacity by adding additional blades; a complete Satori Server environment can scale to support hundreds of terabytes, Hinshaw says.

Just as important, organizations can add additional server blades without rebuilding racks or rearchitecting their databases or applications. Each Satori Server blade boasts self-healing capabilities—i.e., RAID-5 support, along with hot spares, mirroring, and record-level multi-versioning.

Data warehousing veteran Mark Madsen, a principal with consultancy Third Nature, says he’s impressed by Dataupia’s appliance offering—although he stresses that he hasn’t had a chance to put it through its paces.

"I think the concept has a better chance to succeed than the traditional database appliances because Dataupia operates a level lower. You get a reduction in cost since the heavy lifting is done in the storage layer and you need fewer CPUs under the database, but you still get to keep the database you're familiar with," he points out. "Without hands-on experience it's hard to judge, but I would say it will be a lot simpler to grow a database using their hardware since you aren't worrying about operating system tuning, storage subsystem tuning, adding to your platform, or dealing with database options."

At the same time, Madsen allows, the Dataupia appliance, like those of its competitors, flies in the face of a staunchly entrenched trend. "The problem with both appliances and alternative databases is that they go against the current of standardizing to one database provider. Even though they are logically equivalent black boxes you point a JDBC driver at, they are something different and IT usually views different as bad," he indicates.

Nevertheless, Madsen concludes, there’s definitely a market for appliances—as Netezza and DATAllegro have both demonstrated. "For people at the high end, new technologies to cope are always going to be attractive because the problems are so expensive to solve and there's value in solving them. I don't think the traditional database and hardware vendors have done a good job reducing the complexity of large database installations. That's why we have all these new contenders in the market. There are a lot too, from the traditional appliance guys to new entrants like Paraccel and old new entrants like Vertica," he concludes.