Big data-as-a-service player Qubole is getting up to speed with the addition of Apache Spark to its Qubole Data Service (QDS) platform. The company, which was founded by the team that developed Facebook’s data infrastructure, claims that the addition of Spark broadens the types of workloads that can be performed on QDS via the cloud, without the hassles, costs and risks of deploying Spark on-premise.

Qubole launched QDS in June 2012 and opened its platform for public consumption in December of that year. QDS can be thought of as a “self-service” platform that’s designed to lets users run a variety Hadoop jobs — including Hive, MapReduce and Pig — on major public clouds like Amazon AWS, Google Compute Engine and Microsoft Azure. It’s essentially a turnkey solution in the cloud that automatically provisions clusters and scales them according to usage and use case.

According to Qubole, the approach leads to significant cost savings, solves the “skill shortage” dilemma, and reduces the risk of deploying Big Data. One of the biggest advantages is the platform’s ability to automatically set up and scale clusters to match the needs of a particular job, before widning down nodes when they’re no longe required.

With the new update, Qubole’s users will be able to gain access to Spark’s rapid in-memory processing capabilities that make it the ideal choice for running predictive analytics and machine learning jobs. Qubole claims its possible to have a Spark cluster up and running in just 15 minutes or less, saying that the platform makes it easy to set up multiple user accounts and to launch and configure multiple Spark clusters as needed. Spark-as-a-Service has adopted the standard Spark notebook-style interface, which means commands can be made in the Scala, Python and R programming languages, while the QDS platform provides inline results and template visualizations for Spark queries.

“By adding Spark to QDS, we’ve completely eliminated the barriers to taking full advantage of Spark for rapid data analytics and we’re giving customers the ability to select the best technology for their big data tasks, on the fly,” said Joydeep Sen Sarma, co-founder and CTO of Qubole.

Qubole says a number of its customers have already deployed Spark on QDS, including DataLogix and Pinterest.

photo credit: unfetteredhowl via photopin cc;