Big DataCloudCloud ManagementData StorageDataBaseSoftwareWorkspace

Cloudera-Powered Big Data Appliance Announced By Oracle

Editor of eWEEK and repository of knowledge on storage, amongst other things

Follow on: Google +

Oracle has launched the Big Data Appliance which was forged from a partnership with Cloudera, the Apache Hadoop interface/tools provider

Even though Oracle launched  its 12-petabyte data-storage Exadata server in 2008 and its cloud database/analytics Exalogic server to equally high visibility two years later, it still believes it needs a product that is actually named “big data” so there is no mistaking its purpose.

The company has launched a data centre machine called the Oracle Big Data Appliance, which follows in the footsteps of many Oracle products: Engineered-together hardware and software, previously tested and configured, and stamped with the Oracle and partners “seal of approval” (though not literally). It is all but a turnkey workhorse to provide analyses of large batches of data.

Hadoop in the loop

Oracle announced that it has forged a partnership with pioneering Hadoop analytics software provider Cloudera to create a specific Apache Hadoop distribution and tools for the Big Data Appliance.

“A good number of our customers want to just ‘load and go’ when it comes to big-data batch (warehousing and analytics) use cases,” Cetin Ozbutun, Oracle vice president of Data Warehousing Products, told eWEEK. “There is minimal configuration and setup with this appliance, and we’ve found that customers can save weeks or even months on getting their deployments up and running.”

The Big Data Appliance is designed to provision a highly available and scalable system for managing and analysing massive amounts of data in Hadoop and using R (the open-source programming language) on raw data sources, Ozbutun said.

“It also simplifies and controls costs for the big data analytics process by re-integrating all hardware and software components into a single big data package that complements (existing) enterprise data warehouses,” Ozbutun said.

Cloudera’s was the first commercial implementation of the open-source data analytics package that came out of Yahoo’s R&D division in 2006. Cloudera is an equal-opportunity Hadoop developer; last August, the company signed up to provide Oracle.

Hadoop is complicated software to deploy and use, and it lacked a user-friendly front end until Cloudera and others came in to add their interface expertise. Using Cloudera, Oracle’s Big Data Appliance gives users a single source to deploy, manage and scale this new Apache Hadoop-based stack.

Hardware details

On the hardware side, the Big Data Appliance comes in a full-rack configuration of 18 Oracle Sun servers with a total of 864GB main memory; 216 CPU cores; 648TB of raw disk storage; 40Gbps InfiniBand connectivity between nodes and other Oracle engineered systems; and 10G bps Ethernet data centre connectivity.

The new engineered-together system scales by connecting multiple racks together via an InfiniBand network – like Exadata and Exalogic – which enables it to acquire, organise and analyse extreme-size data volumes consisting of machine- and/or human-created data streams, Ozbutun said.

Included in the Big Data Appliance rack is Cloudera Manager and the Oracle NoSQL Database, which can scale horizontally to hundreds of nodes with high availability. The appliance can run both Oracle NoSQL Database Community and Enterprise Editions, Ozbutun said.

The appliance, which was announced at Oracle OpenWorld last September and costs $450,000 (£300,000), includes Oracle’s Linux operating system to run the 18 Oracle Sun servers. Oracle’s Big Data Connectors to other data warehouse cost $2,000 (£1,293) per processor licence.