Date of Award

Fall 12-2013

Degree Type

Honors College Thesis

Department

Computing

First Advisor

Shaoen Wu

Advisor Department

Computing

Abstract

Cloud computing is a “new frontier” in the world of computing. One of the cloud architectures widely used is the Hadoop running environment. Hadoop consists of many parts—including MapReduce, TaskTrackers, and JobTrackers. Right now, there is no fault-tolerance for JobTrackers in Hadoop. This paper analyzes four different distributed hash algorithms (Pastry, Tapestry, CAN, and Chord) that could be implemented inside Hadoop to improve JobTracker fault-tolerance. We recommend Chord as the best suited for integration and improvement of Hadoop.

Share

COinS