theCUBE Live At Hadoop World 2014

The characteristically active analytics ecosystem went into overdrive for last week’s landmark Strata Conference + Hadoop World 2014 summit in New York to produce some of the most significant product updates in recent memory.

Cloudera set the tone for the event the evening before the official kickoff with a major update to its flagship distribution of the batch processing framework that taps into the elasticity of the public cloud to make it easier to accommodate the growing amounts of unstructured information flooding into the enterprise.

Cloudera Enterprise 2.2 achieves that with a new management service that provides a centralized interface for running Hadoop on and across multiple cloud environments, although only Amazon Inc.’s market-leading infrastructure-as-a-service platform is supported on launch. The console also includes capabilities for monitoring and securing off-premise analytics clusters, taking advantage of the tighter controls added to several of the core components included in the company’s distribution as part of the update.

Cloudera’s two main competitors in the pure-play Hadoop field weren’t standing by idly. MapR said it will bundle its powerful core NoSQL database engine into a free Community Edition distribution of its Hadoop bundle. The company dismissed questions about whether the move would cannibalize its business, saying the seeding strategy would expose more customers to its technology.

Hortonworks focused on the ecosystem surrounding Hadoop with its announcements, adding support for the Apache Spark query engine and Apache Kafka message broker and committing to a project to add object database support to Hadoop.

Perhaps the biggest surprise announcement at the show was from Cray Research. The venerable supercomputer maker debuted an integrated Hadoop appliance that sports the kind of huge processing and memory capacity for which Cray is famous.

Cloudera is hardly alone in wanting to harness the scalability of the cloud to address the unstructured data explosion. Microsoft is aiming for the same goal with the addition of support for Apache Storm to HDInsight, a version of Cloudera rival Hortonworks Inc.’s Hadoop version adapted to run on its Azure cloud, which in turn competes with AWS.

Accepted as a top-level project in the Apache Software Foundation earlier this month, Storm provides a scalable and reliable way to ingest fast-moving streams of unstructured data such as tweets and live video. The integration, which is currently in preview, is Microsoft’s attempt to bring HDInsight up to par with Amazon’s Kinesis service, which is designed for the same use case and which already supports Storm. To empower developers to take advantage of the project, the Redmond giant is also rolling out a suite of per-packaged machine learning capabilities for tapping into real-time data.

A startup called Loggly Inc., meanwhile, ​ raised $15 million in fresh funding to help its users drill into a key subset of that information: the automated logs coming off organizations’ infrastructure and applications. It marked the round, which brings its total raised to over $30 million, with the release of a new overlay that eliminates the need for engineers to fret over spelling and scan individual events in order to understand why a particular system is behaving in a certain way.