Microsoft has expanded on its big data cloud services with Azure Data Lake Analytics, a new toolset to be released as a preview later this year along with Azure Data Lake Store. The firm also announced the availability of Azure HDInsight running on Ubuntu Linux.
Microsoft announced Azure Data Lake at the Build developer conference earlier this year, describing it as a highly scalable service that would enable customers to pool and analyse large volumes of data from multiple sources, such as sensor data from the Internet of Things.
The company has now announced that Azure Data Lake will be available in preview later this year, rebadged as Azure Data Lake Store. Also available as a preview will be Azure Data Lake Analytics, the two forming part of a new and expanded Azure Data Lake platform that makes big data processing and analytics simpler and more accessible, the firm said.
Azure Data Lake Analytics is a new service built on the Apache Yarn cluster management tool. It scales dynamically so that customers can focus on their business goals, not on managing distributed infrastructure, according to T K Rengarajan, corporate vice president for data platforms at Microsoft.
"The analytics service for Azure Data Lake is cost-efficient because you only pay for your job when it is running, and support for Azure Active Directory lets you manage access and roles simply and integrates with your on-premises identity system," he wrote on Microsoft's SQL Server blog.
Azure Data Lake Store provides a single repository for big data, which customers can use to capture data of any size, type and speed without forcing changes to the application code as data scales. It is designed to underpin Azure Data Lake Analytics as its storage layer (see image), as well as other big data tools such as the Hadoop distributions from Cloudera, Hortonworks and MapR.
Azure Data Lake Analytics includes Microsoft's Azure HDInsight tool, which is based on the Hortonworks Data Platform, and Microsoft announced that the latter is now available running on Ubuntu Linux as well as Windows Server clusters.
The move was first detailed at the Strata + Hadoop World conference in February, when Microsoft made it available as a public preview. The firm said at the time that it should prove compelling for customers already running Hadoop on Linux, as it would make it possible to use common Linux tools to extend deployment to Azure via a hybrid cloud connection.
This was echoed today by Canonical's John Zannos, vice president of Cloud Alliances.
"The collaboration between Microsoft and Canonical to create the option to run Azure HDInsight workloads on Ubuntu or Windows gives Azure customers flexibility in their big data processing decisions, both on premise and in the cloud," he said, writing on the firm's blog.
More fingers of blame pointed at gangs linked to North Korean government
Dominance of Apple and Samsung in smartphones being chipped away by Huawei, Oppo and other cheaper rivals
OLED smartphone display can be stretched, bent, rolled and even dented - but won't break
Upgrading from a conventional hard-disk drive to an SSD? This may be just what you're looking for