The Fifth Elephant 2012

A discussion on the evolution of bigdata systems within InMobi and a discussion/demo of an in-house data processing and analytics system for large data at InMobi scale.

Outline

Data processing, analysis and visualization of data at a scale at which web systems (like InMobi) operate, is a very hard problem. I’ll discuss the evolution of the data analytics at InMobi - the approaches that we tried, the challenges faced, and our rationale to develop an in-house analytics system on top of Hadoop. I will be discussing the details of this system (moderate level), how it is helping us attain better efficiency levels and what future directions could it take. I will also be doing a demo of the capabilities of this system towards the end of the talk.

Requirements

Basic familiarity with the concepts of data analytics and Hadoop.

Speaker bio

Gaurav Agarwal currently leads the Data Analytics system at InMobi. He has 5+ years of experience building analytics for extremely large amounts of data at InMobi and, previously, at Google. He holds a Masters degree in Computer Science from University of California.

Links

• http://www.slideshare.net/GauravAgarwal35/yoda-fifth-elephant