Netezza Turns a Corner
If the mantra of business today is "faster, better, cheaper," then it's inevitable that companies need to explore alternatives to traditional enterprise software.
- By Wayne Eckerson
- October 14, 2009
Editor's note: This article previously appeared (in a slightly different form) at Wayne's World, TDWI Research director Wayne Eckerson's blog.
Once compared to David facing Goliath, database vendor Netezza yesterday traded in its wooden slingshot for steel blades and armor as it both celebrated its victories to date and geared up to fight bigger adversaries.
With nearly 300 customers and a new "commodity-based" architecture, Netezza is the clear leader in the emerging analytics database market that it evangelized beginning in 2002. It celebrated is stature and market position with a high-energy, one-day event in Boston that kicks off a worldwide roadshow.
Almost 400 attendees (a good percentage being Netezza employees to be fair) heard CEO Jim Baum introduce the new Netezza TwinFin data warehouse and analytic appliance that boasts higher performance, greater modularity, and more openness and flexibility to handle various workloads than its predecessor boxes. Netezza primarily designed TwinFin to address new customer requirements but secondarily to strike a blow to "new" arch competitors Teradata and Oracle.
"We're 50 times faster than Oracle and ten times faster than Teradata without tuning," said Baum. Other Netezza executives added that the company's platform is no longer a one-trick pony. "We're moving from supporting fast scans to multiple workloads that address broader market demand," said Phil Francisco, VP of product management and product marketing, in a pre-briefing this summer.
While none of the customers it highlighted during the day-long event on Boston's waterfront is using TwinFin in production, most touted dramatic performance gains compared to competitor products in head-to-head tests. Some were testing TwinFin and most seemed eager to take the plunge, especially high-end customers who could benefit from the open architecture that makes it easier to embed advanced analytics programs.
Migration and Markets
What wasn't discussed, however, was the time and cost to migrate from Netezza's existing boxes to TwinFin. I raised this question with Netezza marketing cohorts Phil Francisco and Tim Young this summer. They said that most existing Netezza customers will eventually swap out old boxes with TwinFin as they exceed the capacity or performance requirements of existing machines. One of the limitations of the old architecture was relatively fixed capacity boxes with limited extensibility for growth. They implied that this type of forklift migration would be no surprise to existing customers, which is one of the downsides of selling "appliances" where everything is pre-built inside a box.
However, that doesn't mean that Netezza will stop selling appliances—by definition, a complete hardware/software solution that requires no assembly and minimal configuration and tuning. You won't be seeing a software-only version of Netezza TwinFin for sale. Netezza will still ship out appliance machines, except now the new open architecture dramatically improves the system's overall flexibility and extensibility.
Perhaps more importantly, TwinFin is designed to address a new market. "We've moved past the early adopter, risk-taker market and are addressing a more mainstream customer," said Young. While existing customers will be the first to adopt TwinFin, ultimately the new product is designed to position Netezza as an attractive alternative to leading database vendors.
Still, it takes some courage to revamp your architecture when you have a large installed base. But many young, ambitious companies have had to undergo this painful baptism—sometimes multiple times—to position themselves for future growth. Business Objects, MicroStrategy, and Cognos all come to mind.
But Netezza may get off easy. No one seemed up in arms about the architectural shift, probably because most people—customers, analysts, press—seem to understand the benefits of building database systems on commodity hardware. (We should thank Greenplum, Aster, and other analytic database upstarts for that.) The shift means that it will spend less time engineering hardware and more time delivering value-add capabilities through software.
Netezza already is attracting many partners who find it easier to build applications and plug-ins within the open architecture running on blade-based servers, commodity disk storage, Intel chips, Linux operating system, and SQL, ODBC, JDBC, and OLE DB data access interfaces. Also, Netezza will enable customers to deploy TwinFin appliances in hub-spoke configurations with up to 10 racks to increase extensibility. It will also soon introduce specialized TwinFin appliances tuned for different workloads, including high capacity storage (up to 5 PB) and extreme performance (1 TB RAM per rack), a move eerily similar to Teradata's lineup of new appliance machines.
Despite the hype, the future of TwinFin rests in the hands of customers who will vote with their pocketbooks. Netezza has claimed some impressive victories to date, but it has many battles to fight before it can claim the crown.