This week Cloudera, a provider of software and services for the Apache Hadoop project, is set to announce a partnership with Quest Software to develop, support, and distribute an Oracle connector for Hadoop.
Hadoop is the popular open-source implementation of MapReduce, a powerful tool designed for deep analysis and transformation of very large data sets. It enables its users to explore complex data, using custom analyses tailored to users' information and questions.
Code-named "Ora-Oop," the connector will provide connectivity between Cloudera's Hadoop distribution and Oracle through an interface that allows for bidirectional, scalable, and functional data transfer between Oracle databases and Hadoop. Ora-Oop is designed to compliment Sqoop, an open-source tool packaged within CDH designed to import data from relational databases into Hadoop.
Billy Bosworth, vice president and general manager for Enterprise Database at Quest Software, told CNET that this is Quest's first foray into developing an open-source component in concert with Cloudera and that Quest chose to work with Cloudera based on Cloudera's expertise in Hadoop as well as its understanding of how RDBMS and NoSQL databases can play well together.
NoSQL tools are a set of operational-data technologies based on nonrelational technology. These technologies do not always replace the relational database but rather add a new tool to the developer toolbox.
Quest comes from a relational database management system (RDBMS) perspective and has a very strong background in Oracle tooling. Over the last year the company noticed an opportunity in its customer base to introduce new projects that relied on databases and methods associated with NoSQL projects.
Few "classical RDBMS" customers are familiar with the NoSQL movement, according to Christian Hasker, Quest director of product management. According to Hasker, most customers are looking to solve specific-use cases but tend to be very database oriented, not hardcore Java programmers.
Hasker suggested that in order to increase enterprise adoption, DBAs needed to get more comfortable and have tools and mechanisms in place that allow people to continue to manage databases rather than having to learn new development methods. NoSQL databases and projects like Hadoop are often developer driven, and developers tend to be more bleeding-edge than DBAs.
Quest is tracking a number of NoSQL databases and database approaches, including Cassandra, HBase, Amazon SimpleDB, Microsoft Azure, and databases in the cloud, and the company will add more as it learns what the community wants to see.
One interesting note is that Oracle has thus far not paid much attention to NoSQL and has been dismissive of most NoSQL databases, much in the same way the company had been in regard to cloud computing. But, as the company has tended to in the past, Oracle will likely eventually embrace NoSQL and probably release its own flavor of some NoSQL-like product, including the possibility of its own MapReduce derivative.