Appliances for Business Intelligence
TDWI's Director of Research examines Netezza Performance Server
A sure sign of the maturity of any technology market is the advent of plug-and-play “appliances” that pre-integrate components to deliver a cost-effective data center system that is easy to set up, use, and mange.
Netezza, a Framingham, Massachusetts-based start-up is betting that the business intelligence market has matured to the point where organizations will see the logic in swapping expensive data center platforms consisting of separate storage, servers, and database systems for its “tera-scale data appliance” that combines off-the-shelf versions of these components in a single, rack-mountable cabinet.
The Netezza Performance Server is geared to support complex analytic queries against large volumes of detailed data. It delivers 10 to 20 times the performance at half the cost in these types of environments, according to company officials. The Netezza 8100 Performance Server, which pre-integrates hardware, storage, and a SQL database, costs $622,000 and installs in less than one day.
Not surprising, there has been significant interest in the BI appliance from organizations that have not been able to justify the high costs of loading, storing, and querying large volumes of detailed operational data. One significant market for Netezza is analytic service providers who perform complex analytics on client data.
“We manage and analyze large databases for our clients and the Netezza Performance Server has shown dramatic performance improvements over our existing systems,” says Mike Coakley, a vice president at Epsilon, a database marketing services provider. Epsilon is consolidating numerous client-specific data marts onto a single Netezza BI appliance. “Queries are running eight to 200 times faster and we can now refresh a data warehouse with tens of millions of records in 15 minutes.”
Other target markets for Netezza include telecommunications firms that want to analyze call detail records; financial services firms that need to analyze risk, fraud, or credit card trends; and retail firms that would like to perform detailed market basket analysis, among others.
Netezza lowers total cost of ownership because it uses off-the-shelf components, such as Linux servers, gigabyte Ethernet switches, Pentium processors, Compaq servers, IBM disks, and an open source database it has modified to work with its architecture. The system provides fast query performance because it performs many common query-processing tasks in specialized computer chips located on arrays of massively parallel disk drives within its storage subsystem. Netezza’s patent-pending technology essentially blends the best of both SMP and MPP processing to optimize query performance against large volumes of data.
Since it supports SQL, ODBC, and JDBC, the Netezza appliance works with most analytical tools on the market today. To date, Netezza has established formal partnerships with Microstrategy, Business Objects, and SPSS, with others to come.
Netezza officials understand that most organizations are loathe to change their internal BI platforms, and thus they often face an uphill battle to win customers. However, they say organizations are eager to hear their story when they face a costly upgrade to their hardware, storage, or database system or need to create a BI infrastructure from scratch in a new project.
To evangelize its BI appliance, Netezza performs proof-of-concept (POC) demonstrations free of charge. Netezza will gather customer data and queries and test them on systems in its labs. Customers can then bring the system in house to validate the lab results in their own environment. The only caveat in the validation phase is that customers must commit to buying the system if it matches the laboratory performance.
Currently, Netezza has three paying customers, including Epsilon and a telecommunications services firm. It has shipped systems to two other prospects and is conducting numerous POCs. If Netezza is right and the BI market is mature, expect this start-up to have a dozen customers by mid-year and several imitators ready to ship BI appliances in the near future.
For more information, see http://www.netezza.com/
Wayne Eckerson is director of research at The Data Warehousing Institute (TDWI), a provider of in-depth education and research in the business intelligence and data warehousing industry.