Nimbula and MapR Tackle Private Cloud-Based Hadoop
- By Dian Schaffhauser
- 08/14/12
A company that offers a cloud platform has teamed up with another that distributes a version of Apache Hadoop to enable customers to expedite their cloud implementations of Hadoop clusters. The joint solution combines Nimbula Director with MapR Technologies' Hadoop distribution. It includes templates, recipes, and verification tests for running Hadoop on a Nimbula Director cloud.
Nimbula Director is a private cloud operating system that turns "bare metal servers with local disks" into multi-tenant pools of computing resources with a self-service interface for provisioning and de-provisioning of Hadoop and other jobs.
Hadoop is an open source framework that allows for the distributed processing of large data sets across clusters of computers. It can work on a single server or thousands of servers, each offering local computation and storage. Because it's architected to be highly available, Hadoop overcomes individual machine failures that can surface in the use of commodity hardware.
It's also commonly used in higher ed computing centers. For example, Stony Brook University uses Hadoop to process and extract statistics from a thousand-plus daily newspapers as well as historical archives. The University of Maryland is using its implementations to do research in machine translation, language modeling, bioinformatics, email analysis, and image processing. The University of Nebraska-Lincoln's Holland Computing Center is using Hadoop in student research projects.
With the latest announcement from Nimbula and MapR, customers can provision Hadoop clusters, according to the companies, in under two minutes. Users can launch Hadoop clusters on a private cloud without having to put in place the underlying computing infrastructure. The integration of technologies allows for sharing of infrastructure between Hadoop clusters with permissions as well as network and resource isolation. Hadoop and non-Hadoop workloads can run on the same infrastructure.
"With this solution, customers can have the best of two worlds. They can have big data processing from Hadoop with private cloud's ability to deliver low cost shared infrastructure that manages elastic demand between multiple tenants," said Jay Judkowitz, director of product marketing at Nimbula.
Nimbula Director is available for free for deployment of up to 40 cores. Customers can download a packaged VM template and application definition, load it into their Nimbula Director cloud, launch it, use it as long as they need, and remove it once their jobs are completed. That offering can be downloaded with registration from nimbula.com.
About the Author
Dian Schaffhauser is a former senior contributing editor for 1105 Media's education publications THE Journal, Campus Technology and Spaces4Learning.