The drug market-analysis team at Wolters Kluwer Health used to turn away customers every month. It wasn’t because they didn’t have enough staff to handle every data request from customers in the pharmaceutical industry. The problem was their Oracle-based data warehouse couldn’t handle all the requests quickly enough.
The drug market-analysis team at Wolters Kluwer Health used to turn away customers every month. It wasn’t because they didn’t have enough staff to handle every data request from customers in the pharmaceutical industry. The problem was that their Oracle-based data warehouse couldn’t handle all the requests quickly enough.
“Because of the processing times to perform the queries in Oracle, we were literally capacity-bound in terms of . . . the amount of work we were able to get out of the door in a month,” says Emory Heisler, vice president of global IT services for the Wolters Kluwer healthcare analytics team in Phoenix. “We had to turn work away.”
This changed about a year ago when the healthcare analytics division switched a portion of its data warehousing from Oracle to Netezza.
Netezza offers a data warehouse appliance that bundles hardware, software and storage capacity in one prepackaged unit. Complex data-analysis queries that used to take days now take minutes, Heisler says.
Netezza says the unique aspect of its system is the placement of processing power right next to the data storage, which eliminates the need to shuttle records sets back and forth between storage systems and processors.
Wolters Kluwer’s healthcare analytics team, which provides drug-utilization data to companies like Pfizer and GlaxoSmithKline, now stores 9T bytes of data on a Netezza system, and about 40T bytes on an Oracle system. The division continues to use Oracle for queries that are not particularly time-sensitive, but Heisler says productivity has increased 25% in the portion of the company that is using Netezza. In addition, the company has saved money because the Netezza database doesn't require as much internal staff to maintain, he says. For users, the simplicity of the data warehouse appliance is “eye-popping,” he adds.
Netezza claims its product lets data queries be performed 10 to 100 times faster than systems like the one made by Oracle. In some cases, this is true, says analyst Richard Winter, president of research and consulting firm Winter Corp. in Waltham, Mass. But Oracle’s data warehousing product is more mature and better able to handle the most complex database structures, says Winter, who has consulted for both Oracle and Netezza.
The products made by Oracle and Netezza represent two ends of the data warehousing spectrum, Winter says. Oracle, like competitors Teradata and Microsoft’s SQL Server, sells data warehouse software and lets customers choose their own server, storage unit and operating system, he says. This gives customers the freedom to make changes to their data warehousing infrastructure, and it also makes it easy to incorporate Oracle into an existing system.
“Almost no matter what your environment, you can bring Oracle into it and you can use the standards you’ve established in your environment. You don’t have to change anything,” Winter says.
On the other hand, companies that use Oracle have to buy from several suppliers to get all the necessary components.
“With the Netezza approach, it’s a completely integrated solution,” Winter says. “They’ve made all those choices. They’ve decided what’s going to be the server, what’s going to be the storage, what’s going to be the operating system. What you get is a box, you get an appliance.”
Convenience and cost are the major advantages. With Netezza and other data warehouse appliances, a customer needs little or no staff to maintain the system, Winter says.
Winter says he and his colleagues have interviewed three Netezza customers who confirmed that their data queries became 10 to 100 times faster when they switched to Netezza. That company is still relatively new, however, and may not be mature enough to handle the most complicated database needs, he says.
“The cautionary note is these [data warehousing appliances] are not mature products, and they have to be evaluated carefully,” Winter says.
While Wolters Kluwer deals with “massive volumes of data,” the actual database schema -- the structure -- is relatively simple, Heisler says.
A Netezza data warehouse appliance can cost as little as $200,000, or in the millions of dollars for high-end systems used on a large scale, says Jim Baum, Netezza’s president and COO. A new customer can install a typical rack with 12.5T bytes of capacity, load data and begin doing queries within hours, he says.
Since starting up in 2003, Netezza has acquired 95 customers, including Amazon.com, the American Red Cross and the U.S. Navy, he says.
Heisler says he would recommend Netezza to companies performing high-volume analyses of large amounts of data. The appliance performs well when data remains static, changing little throughout the day, he says. Netezza would not be the best device to assist in the processing of online credit card transactions, but would be good for analyzing those transactions to discover market trends, he adds.
Despite switching part of its data warehousing capability from Oracle to Netezza, Heisler says he has no plans to get rid of Oracle entirely. “Oracle still applies and is a fabulous tool where it fits,” he says. “We don’t intend to convert all of our databases to Netezza.”
Learn more about this topicSOA driving interest in virtual data warehouses
9/27/06Q&A: Appliance boosts data warehouse performance, Netezza exec says
9/15/06Farecast.com uses Oracle data warehouse to predict holiday airfare trends