Twister Collective Iterative MapReduce Development

Abstract

To investigate system issues of large scale data analysis using iterative MapReduce framework, we use Bravo and run Twister on it. With the large memory machines in Bravo, we can test Twister with large input data.

Intellectual Merit

To investigate system issues of large scale data analysis using iterative MapReduce framework.

Broader Impact

To accelerate iterative computation that fully utilizes large memory machines.

Use of FutureGrid

Deploy Twister on Bravo nodes and run K-means Clustering application.

Scale Of Use

All nodes in Bravo

Publications


FG-275
Bingjing Zhang
Indiana University
Active

Keywords

Bravo, mapreduce, twister

Timeline

2 years 5 days ago