The Channel logo


By | Cade Metz 2nd October 2009 10:02

Web GUI hides number-crunching open source elephant

Cloudera's Hadoop desktop

Hadoop World Cloudera - the startup that commercialized the Google-like number crunching of the open source Hadoop project - has unveiled the first web-based GUI for running applications on the much-hyped distributed computing platform.

Announced this morning at the Cloudera-sponsored Hadoop World in midtown Manhattan, the free Cloudera Desktop provides a graphical user interface for creating and submitting jobs on a Hadoop cluster, monitoring the cluster's health, and browsing stored data. Typically, Hadoop clusters are managed via the command line.

"We always knew we had to create some kind of Hadoop management console," Cloudera vp of product Jeff Hammerbacher tells The Reg. "The idea is to make Hadoop more accessible and usable by non-developers and non-web-organizations."

Inspired by research papers describing Google’s proprietary software infrastructure, including its GFS file system and MapReduce platform, Hadoop is a means of crunching extremely large amounts of data across a network of distributed machines. The Apache-hosted project was founded by Nutch crawler creator Doug Cutting, who spent several years driving Yahoo!'s Hadoop work before leaving for Cloudera. Famously, Cutting named the project after his son's yellow stuffed elephant.

The Cloudera Desktop runs insides a browser, but it's meant to have the look and feel of a desktop OS. Make that Apple's Mac OS X:

Cloudera Desktop

(Early incarnation of) the Cloudera Desktop

The initial version of the pseudo-desktop environment includes four applications: a file browser for copying and browsing data files; a job submission tool for creating jobs, running them, and saving them for later use; a job browser for tracking live jobs; and a health dashboard for monitoring the cluster and alerting users to problems.

Unlike the Hadoop platform itself, the Cloudera Desktop is not open source - at least not now. And for the moment, it only works with Cloudera's Hadoop distro - though Hammerbacher says it will eventually front other distros as well.

The ex-Facebooker also says that the company has pieced together a developer toolkit that should soon deliver additional applications for its Desktop. These include an app for writing Hive queries, Hadoop's SQL-like coding language, and a GUI version of SQOOP, the Cloudera-created tool that imports data from existing databases. For its own internal Hadoop implementation, Facebook offers its semi-technical employees a Hive GUI known as HiPal

The Cloudera Desktop requires the installation of a server, which can be downloaded here. At some point, Hammerbacher indicated, third-party developers will also have access to the tool-kit. ®

comment icon Read 1 comment on this article alert Send corrections


Frank Jennings

What do you do? Use manual typwriters or live in a Scottish croft? Our man advises
A rusty petrol pump at an abandoned gas station. Pic by Silvia B. Jakiello via shutterstock

Trevor Pott

Among other things, Active Directory needs an overhaul
Baby looks taken aback/shocked/affronted. Photo by Shutterstock

Kat Hall

Plans for 2 million FTTP connections in next four years 'not enough'
Microsoft CEO Satya Nadella


Suit-and-tie-wearing man tries to meditate, take deep breaths in faux yoga pose. Photo by Shutterstock
Emotional intelligence, not tech skills, is the way to woo suits
League of gentlemen poster - Tubbs and Edward at the local shop. Copyright BBC
One reselling man tells his tale of woe