Original URL: http://www.theregister.co.uk/2010/10/12/cloudera_membase/

Cloudera and Membase unveil NoSQL togetherness

Hyper-scale analytics

By Gavin Clarke

Posted in Applications, 12th October 2010 23:26 GMT

Cloudera is deepening its relationship with NoSQL partner Membase to jointly attack customers running the open-source implementation of Google's MapReduce.

Membase, until recently called NorthScale, has announced software and services to help customers running the MapReducesque Hadoop with its key-value database, called Membase Server. Cloudera is the commercial venture trying to sell services in support of Hadoop.

The start-up has announced the NodeCode module to stream data from its key-value database store Membase Server to Cloudera Distribution for Hadoop (CDH) in real-time, along with a batch-load utility based on Sqoop to transfer data between Membase and CDH.

Membase said it has also devised a program to sell Membase and Hadoop together based on specific customer use cases. Those cases are advertising, offer and content targeting, log and event stream capture and analysis, and social gaming.

Illustrating the fact that Cloudera and Membase engage on joint customers, Membase announced AOL Advertising and ShareThis have built what it called advanced ads targeting systems to crunch into user profiles large volumes of event information tied to users via cookies.

ShareThis is the service behind the button on 850,000 web sites that lets users share content via email and social networks. AOL, meanwhile, uses Hadoop for ETL, style processing, and to generate stats for behavioral analysis and targeting.

The expansion comes after Cloudera chief executive Mike Olson, a database-industry veteran and former chief of Berkeley DB specialist Sleepycat Software, joined Membase's board earlier this year. Venture fund Accel Partners is an investor in both companies.

Olson said in a statement that integration between CDH and Membase would result in a "highly optimized data delivery system with virtually no lag time."

"This real-time processing capability is essential for any solution on which split decisions must be made, including ad targeting and social gaming," he said.

Membase blogged that profiles would be fed into Membase and served with sub-millisecond latency. This will combine high-powered analytics with raw real-time performance, the company said. ®