In-Depth

Why Oracle’s Exadata May Attract CXOs

• 12/02/2009

We're still waiting for independent corroboration of Oracle Corp.'s Exadata version 2 (V2) performance and pricing claims. That could take awhile -- in the case of Exadata V1 (aka Oracle Database Machine), the first such assessment -- an Oracle-sponsored study conducted by DW analyst Richard Winter -- didn't appear until several months after Oracle's announcement.

In the absence of hard data, there's always speculation, which is why a new Exadata market report from IDC is so intriguing.

First, it was written by two non-BI or DW-practitioners -- namely, Jean Bozman (one of IDC's most esteemed server market watchers) and Matthew Eastwood, who heads up IDC's Enterprise Platform Group.

Neither Bozman nor Eastwood can claim any particular BI or DW market expertise. However, both analysts know Big Systems, are conversant with Big Data problems -- chiefly in connection with online transaction processing (OLTP) workloads -- and have an undeniable feel for the hardware that powers the enterprise.

Hardware isn't sexy, but it is essential. Bozman and Eastwood point out that hardware can play a surprising -- and decisive -- role in the longevity of software assets and the timing and scope of new software purchases.

It's in this last respect that Exadata V2 could prove to be decisive, the IDC pair predicts. Despite the marketing efforts from analytic database players such as Netezza Inc. -- to say nothing of high-end DW stalwart Teradata (or OLTP-oriented upstarts such as Exosol and VoltDB) -- many of the largest data warehousing or OLTP systems in the world continue to run on Oracle.

More to the point, many of the biggest combined DW and OLTP configurations also run on Oracle. Upstart database vendors (whether focused on analytic or OLTP features) can make claims about the performance, price, or elegance of their offerings -- in the Big Data arena, elegance can be an issue. Plenty of DW experts have spoken about the tuning, tweaking, and kludging required to get Oracle (or any conventional DBMS) to scale in the high end, but these shops are staying put, whether from inertia, an attachment to Oracle, or some combination of these is the question.

It's an $8 billion question, according to IDC, which puts the market for data warehousing-related server and storage revenues at just short of that mark.

Oracle is vulnerable in that some of its biggest customers run some version of its database -- in OLTP, DW, or combined OLTP/DW configurations -- on aging SMP servers, Bozman and Eastwood write. These shops are fat targets for Teradata, Netezza, and IBM (which CEO Larry Ellison recognizes as Oracle's biggest DW competitors), along with upstart players such as Aster Data Systems, Greenplum Software Inc., Kognitio (which uses technology from white box DW pioneer White Cross), ParAccel Inc. and Vertica Inc.

With Exadata V2, Oracle is attempting to completely recast the terms of this debate, Bozman and Eastwood say. "In effect, Oracle is making a strong statement about data transformation that could move at least some of the largest Oracle data warehouses in the world, now running on older SMP systems, into a new computing paradigm," they write.

"This is significant because large Oracle databases and Oracle data warehouses are often one key reason that customers prefer to extend their current server hardware, lengthening server lifecycles for scalable servers, especially Unix servers." There's another wrinkle here, too, according to Bozman and Eastwood: with the economy again showing signs of life -- and with customers loosening their purse strings (and in some cases preparing to splurge on long-deferred big-ticket items) -- the pressure's on.

"With the option to move even these largest databases -- ranging from multiple TB to multiple PB in size -- to scale-out x86 server hardware, … then the rate at which these older systems are replaced could increase as the recession lifts," they point out.

Although IDC pegs the market for DW-related servers and storage at about $7.5 billion, it sees a much bigger market -- perhaps as big as $22 or $23 billion -- for data warehousing, OLTP, and analytic workloads. That's huge, and is one reason Oracle's pitch with Exadata V2 -- which struck some DW and BI industry watchers as either superfluous or opportunistic -- is both clever and perhaps even a little audacious.

Quite aside from what data management (DM) or DW pros might think about the desirability -- to say nothing of the practicability -- of combining OLTP and DW onto a single platform, it's a proposition that could appeal strongly to their C-level masters.

"Exadata is now aimed at supporting both Data Warehouse and OLTP workloads, something that would cause it to take on two workloads often combined in scalable Unix servers or datacenter mainframes supporting mission-critical applications," Bozman and Eastwood conclude. "These types of servers are the most visible in the datacenter -- and they are closely watched by senior managers at the CXO level, who want timely business results to increase their enterprise's competitiveness. If queries can be returned in a fraction of the current query times, this will help gain senior management buy-in for any of the systems vendors in the industry that can provide a speedup."