In-Depth
New Competition Validates Importance of Data Warehouse Appliances
Data warehousing firms aren't worried about the rising competition from Oracle, Microsoft, HP, and others.
It's been a time of unprecedented tumult in the data warehousing (DW) market.
Last month, Oracle Corp., in tandem with hardware partner Hewlett-Packard Co. (HP), introduced its most ambitious DW deliverable to date -- the Oracle HP Database Machine. Just last week, Microsoft Corp. outlined plans to develop a version of its SQL Server 2008 database optimized for the requirements of the high-end data warehousing. That deliverable -- slated to ship in the first half of 2010, and dubbed "Project Madison" by Microsoft -- incorporates massively parallel processing (MPP) capabilities from the former DATAllegro Corp., which Redmond purchased nearly three months ago.
What's the reaction from best-of-breed DW players? After all, both Microsoft and Oracle have very deep pockets to challenge Teradata Corp., Netezza Corp., and Dataupia Corp. That's not even taking into account pressure from other heavyweights, including IBM Corp. (which markets its Balanced Configuration Warehouse line of appliance-like systems) and HP (which markets Neoview, an ambitious DW platform of its own).
DW best-of-breeds aren't worried. Take John O'Brien, veteran data warehouse architect and CTO of Dataupia. Oracle's and Microsoft's entries into Big Data Warehousing -- via both reference architectures (Oracle's Optimized Warehouse; Microsoft's Project Madison push) and actual hardware (the Oracle HP Database Machine) -- simultaneously highlights Dataupia's special sauce and heightens the contradictions between his company, its best-of-breed competitors, and the larger competitors.
"What excites me about this is … that it's really an acknowledgement of the value of both MPP and data warehouse appliances. The appliance form factor for any specific technology gives businesses what they want, which is the operational ability to plug-and-play, run-and-go, one-and-done. So they [i.e., businesses] don't waste time and money on the infrastructure and things that aren't core to the business," O'Brien argues. "It's also a pretty good nod to the architecture we put in data warehouse appliances, which is the MPP technology. I think it's a great thing."
Tim Young, vice-president of marketing with Netezza, claims that moves from Oracle and Microsoft merely amount to the mainstreaming of the DW appliance.
"The bottom line is that the data warehouse appliance has gone mainstream. In the space of a couple of weeks, we've gone from being a sort of niche product that people didn't try to understand and that we had to spend a lot of energy educating a lot of people about to something that Oracle and Microsoft are now promoting," he says.
"This isn't a bad thing at all for us. The independents like Netezza are important because we drive innovation, and I still think the market has a long way to go until we reach that point where you get diminishing returns on innovation. It's kind of like what you have in the OLTP [segment], where the change of pace within OLTP databases has kind of massively slowed over the last 10 years."
Big, Fast Still Matter
At the same time, Young says, Oracle's and Microsoft's moves are less about "appliances," per se, and more about the importance of Big/Fast/Scalable data warehousing. "My own sense is that the data warehouse appliance term is actually meaningless. As much as we would like to claim that we invented this term and that we invented this segment, I think what's actually happening is that there's this market emerging -- a market for big, high-end data warehouses that have characteristics for high performance and low latency."
Isn't there a danger that high-end DW deliverables from Oracle and Microsoft could cut into their core markets?
Take Dataupia, whose seminal value-add is that unlike offload-and-replace appliances from Netezza and DATAllegro, its Satori Server is designed to complement existing Oracle and SQL Server implementations.
O'Brien doesn't think so, for a few reasons. First, he says, Oracle's Database Machine (as well as its Optimized Warehouse configurations) are Oracle Database-only plays -- as is Microsoft's Project Madison effort, which is yoked to SQL Server 2008. Secondly, O'Brien argues, neither Oracle nor Microsoft do MPP as effectively as Dataupia or its best-of-breed competitors.
There's evidence to support both contentions. (See accompanying article for a closer look at the MPP features of Oracle and Microsoft products.)
According to one DW architect in a mixed Oracle and SQL Server shop, for example, Oracle's Database Machine is a non-starter for his employer precisely because it's an Oracle-only proposition. He specifically cited Dataupia's Satori Server as a compromise alternative and further noted that his preference would be for a greenfield implementation from Netezza (see http://www.tdwi.org/News/display.aspx?ID=9143).
"[Our] concern is that [the Oracle Database Machine] will obviously only ever work on Oracle systems, but we have several SQL Server [data warehouses] too, so we're also looking into [an appliance from] Dataupia -- which happens to be much cheaper … and [which] works on Sun as opposed to HP [hardware]," this DW architect told BI This Week.
Randy Lea, vice-president of product and services marketing with DW giant Teradata, is likewise unruffled by the new -- and newly ambitious -- data warehousing moves from Oracle and Microsoft. Yes, he concedes, Oracle has a history of aggressive competition, and when Microsoft gets something right, it is a force to be reckoned with. The issue, Lea insists, concerns both the scale of the problem (high-end data warehousing is exploding into petabytes) and the OLTP pedigrees of the big RDBMS vendors.
"This is not a new problem for us. Oracle, DB2, Informix, Sybase: our competitors have always tried to position themselves as a database for everything -- for both OLTP and data warehousing. Quite frankly, the architectures are very different," he argues. "Oracle [and] SQL Server were architected for transactional processing. That's why, to a certain degree, they're trying to add parallelism to an architecture that's transactional related. That [i.e., parallelism] iswhere our tradition is."
The upshot, Lea contends, is that the DW segment is a big (and ever-expanding) pie, and the outsized ambitions of neither Oracle nor Microsoft -- nor HP, IBM, or Sybase, for that matter -- will nibble into Teradata's bottom line.
"Will Oracle delay some motion [toward Teradata] here or there? They might, as customers take some time to try to figure out what it's all about, but every account that we go in we're replacing existing technology. These [data warehousing products] are something that if you're a pure Oracle shop, or you're a pure SQL Server shop, you might be interested in, but for customers that aren't tied to any one platform, they won't be a serious alternative."