"Our goal in 2010 is to demonstrate to enterprises who haven't seen Hadoop before how you can get more value out of data already collected in your relational databases -- which you would leave in place -- by combining it with new data types," he said.
While Olson grants that SQL is an easier and more powerful environment for many users today, he says Hadoop will soon catch up because they "are innovating much faster."
"Why don't we see how long it takes for Oracle to make another major release?" he said.
Hadoop is better at crunching disparate data types than relational-based data marts or data warehouses, which force you to create a schema for the data upfront.
So also, argues Olson, is Hadoop's scalability, saying there are a number of Hadoop clusters storing data "well-known to be multiple petabytes in size." He declined to name who those companies are and whether they are Cloudera customers.
Despite the potential of the Hadoop technology to serve as a scalable, universal data store, Olson sees it complementing, not competing with, relational databases.
"It kinda sucked to compete with Larry Ellison," said Olson, referring to his former firm, SleepyCat Software, embedded database maker BerkeleyDB, which was acquired by Oracle in 2006. "I finally managed to sell the guy a company. So I don't want to [compete with] him again."
Cloudera also works closely with Vertica Systems to enable users to connect data stored in Vertica's SQL-based data warehouse with Cloudera, and vice-versa.
"What Aster Data and Greenplum have is not MapReduce in my view...it's tied only to relational data, not general data," he said. "The reason you would choose Greenplum [MapReduce] is because you'd already be a Greenplum customer, not because you wanted MapReduce."
Eric Lai covers Windows and Linux, desktop applications, databases and business intelligence for Computerworld . Follow Eric on Twitter at @ericylai , send e-mail to firstname.lastname@example.org or subscribe to Eric's RSS feed .