Things have changed since Gartner last published its Magic Quadrant for data warehouse and database management...
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
systems (DBMS) two years ago, an eternity in technology terms.
The 2006 version, released this month, reveals some telling trends, according to study author Mark Beyer, research director with the Stamford, Conn.-based analyst firm. Thanks to drivers like regulatory compliance and increased tactical use of business intelligence, data warehouses and DBMS systems must accommodate mixed workloads of analytic, operational and transactional functions. "Appliances," or bundled hardware and software configurations designed for specific data analysis tasks, were just emerging two years ago. Now, appliances, which tend to be faster to implement and cost less, are changing the market and putting new pressures on established leaders. And leaders have updated their pricing strategies, Beyer said, creating a better buying environment for customers.
Using Gartner's Magic Quadrant methodology, Beyer examined a host of vendors, arranging them into four categories: leaders, challengers, visionaries and niche players. Vendors are evaluated based on a combination of technical vision and ability to execute at customer sites.
This year's challengers, which get high marks for execution but lack vision, were Sybase and Microsoft. The software giant straddled the leadership line, Beyer said, and it may emerge as a leader, pending more data about SQL Server 2005.
"SQL Server 2005 is a good challenging product. Microsoft, of course, can execute, but now we need to wait and see if it lives up to its vision," Beyer said.
The quadrant's visionaries have notable technology but less proof of execution, while niche players are worthy of note but need improvement in both areas. The niche quadrant included U.K-based Kognitio Ltd.; Swedish MySQL AB; and Westmount, Quebec-based Sand Technology Inc., none of which appeared in the study two years ago. The sole niche player two years ago, Framingham, Mass.-based Netezza Corp., has made it to the visionaries quadrant, along with Aliso Viejo, Calif.-based Datallegro Inc.
The visionaries have notably increased pressures on the leaders with their lower-priced appliances, Beyer said. But it's not that they have overwhelming market share, he explained. Rather, the smaller vendors proved that there's a market for preconfigured systems built with commodity-class hardware and open source databases. This has not gone unnoticed.
"Physics" helps appliance market heat up
IBM has built its own appliance. Oracle has joined with hardware partners to create "approved configurations," which similarly lower implementation costs, Beyer said. Teradata has always had an appliance-style, preconfigured system, albeit at the higher end of the cost scale. Whether appliances will rule the market remains to be seen, he said, but they've brought attention to the important issue of the "physics" of data warehouse and DBMS systems.
"The mixed workload of data warehouses means it's critical now to balance the physics of simply moving data through an electronic and digital environment, along with having a good design for your database," Beyer said.
The study identified different kinds of workloads or functions that a system may address, such as enterprise reporting, strategic data mining, tactical business intelligence, continuous data loading, batch loading and ad hoc queries. Enterprises usually require systems that can handle mixed workloads or different combinations of these functions, Beyer explained. This means that companies must choose hardware and software combinations that are appropriately configured and balanced for the required tasks. Whether they choose an appliance or full-blown system depends on their requirements.
Buying advice and other considerations
Traditional data warehouse and DBMS systems are highly configurable, so organizations can tune them for their unique needs. Appliances are generally purpose-built to address specific kinds of workloads, Beyer said. For example, an appliance for strategic analytics or data mining is configured differently than an appliance designed for operational, tactical business intelligence. Try to stretch the use of an appliance, and it probably won't perform as well.
"If you buy an appliance, use it for what it's intended," Beyer advised. "Don't use a blender to try and bake a cake."
Those purchasing a DBMS should understand that they're only buying software, and there are important physics and hardware issues to deal with as well, he said.