NCR Exits the Data Warehouse Market: See You Later, Teradater?

TalkBack 1 of 5:

Next »

• Thread View
• Flat View

Study data warehouses more before slamming market leaders

This BLog is a 100% inaccurate prediction with no substantiating facts. Sun has tried unsuccessfully multiple times to get into DW. GreenPlum is one of a long series of start ups that has a 5% chance of overtaking any one of the leaders: Teradata, DB2, Oracle, and SQL Server. What's different? If price was all that mattered, the Lilliputians would have wiped out Teradata & IBM in DW long ago. But DW technology requires huge investments in RDBMS optimizers, high availability, and people --none of which Greenplum and Sun can muster in less than 5-10 years. The in-memory database technology mentioned is cool stuff - Oracle bought Times Ten to get it. But its a feature of a database, not a database. But in DWs, caching data in memory has a low value because users are always scanning millions of records for sums and analysis, which means you would have to load the entire 10 terabytes into memory before "in memory database" was relevant. This is why adding huge cache memories to disk controllers and DW servers doesn't have the same effect as in other workloads. Go ask IBM why they failed when they installed Shark disk systems on DWs. Don't confuse SAPs in-memory efforts to fix SAP-BW performance problems with a market trend that wipes out incumbent vendors. Finally, Sand Technology has been around for 15-17 years under different names & 3 different companies. Its cool technology but its not a DW killer. Turns out its magic is hampered by the fact that it can take a day to load data into Sand before it can do its magic. That doesn't sit well in companies that are moving from loading data nightly to several times a day. If Sand were going to kill Teradata, Oracle, DB2 or Microsoft, you would be writing obituaries not predictions.

Gartner has full time analysts on the DW market. They have rated Teradata #1 in their annual assessment for 5 years in a row. And they predict the DW market will just keep on growing. Before you frighten stockholders, add this to your Blog. Indeed, you missed the real market threat which is HP's new NEO entry -- but HP intends to sell "largely useless data warehouses" too. Compare this initiative to Sun/Greenplum. The DW market has had 75-100 startups -- both hardware and software -- try to break in and a majority died while some were acquired. Convex, nCube, KSR, Sybase IQ, Redbrick, Informix, Ingres, and the list goes on. Where are they now?

The "largely useless data warehouses" you speak of cost tens of millions. Are you saying you are so much smarter than the 100s of companies & 1000s of people that are willing to pay for such useless things? There are a lot of smart people at Wal-Mart, Citibank, Barclays, Amazon, UBS, and so on. The scrutiny they get every year by the CEO, CFO, and CIO is not easy to survive. If their DW was useless, the budget would be cut. This isn't some mass hallucination. If the DW was useless, you would not be the only one saying it. The Wisdom of Crowds applies.

You might want to explain why the DW market is $40 Billion annually with a growth rate of 5-12% every year since 1995. If Teradata fails after the spin off, it won't be for any of the reasons you cite.

Josh - stick to something you know like SAP.

Posted by: danielg922   Posted on: 01/18/07 You are currently: a Guest | Members login | Terms of Use

Alert moderator to an offensive message

Subscribe to this discussion via Email or RSS

SponsoredWhite Papers, Webcasts, and Downloads

• Improving Server Security Through DRAC 4 User Role-Based Permissions Dell The Dell Remote Access Controller 4 (DRAC 4), introduced in ... Download Now
• Drive Agility & Cost Optimization - NEW SOA and BPM Announcements from IBM IBM Drive Costs Down and Profit Margins Up! Every week, businesses ... Download Now
• Migration From Oracle 9i on Red Hat 2.1 Advanced Server to Oracle 9i on Red Hat Enterprise Linux (RHEL) 3.0 Dell The purpose of this paper is to provide a procedure for migrating a system ... Download Now