Their databases bulging with more than 10 terabytes of raw data, telecommunications companies SBC Communications Inc. and Telstra Corp. topped a survey of the world's largest data warehouses and online transaction-processing databases, respectively. But companies that won high honors in the contest, conducted by market research firm the Winter Corp., say there are neither garlands nor glory in the daily work of maintaining very large databases.
SBC's 10.5-terabyte data warehouse running on an NCR Corp. Teradata system contains marketing, finance, and network infrastructure data. It has doubled in size every year since it was started in 1993. "Obviously, the bigger they become, the more complex they get," says Rolf Hanusa, SBC principal technical architect. Tasks such as database backup and change control become challenges that can only be met by "a combination of good processes and good people to manage those processes," he says.
The Winter survey, results of which were released last week, has become an industry benchmark in defining just how big is big. Among its findings: The number of databases with more than 1 terabyte of data has more than doubled since the last survey in 1998, president Richard Winter says. And 17% of the participating companies expect their databases to increase tenfold during the next three years.
ComScore Networks Inc. isn't surprised. "As fast as I can add capacity, our data warehouse eats it up," says Ric Elert, engineering VP at the Reston, Va., company. ComScore's 2-terabyte data warehouse, which runs on Sybase Inc. software and collects data on Web-surfer behavior, is the largest data warehouse running on Windows NT, the survey found. Elert is adding 143 Gbytes of raw data to the data warehouse every week.
There may even be larger databases out there than the survey reveals. Some companies that operate huge databases and view customer data as a strategic weapon, including Wal-Mart Stores Inc. and United Parcel Service Inc., didn't participate in the voluntary survey this year.