Teaching distributed systems using hadoop

Abstract

Databases and Distributed Systems have a fundamental relevance in Computer Science; they are usually presented in courses where the high-level of abstraction characterizes the teaching and learning processes. Consequently, the teaching method needs to evolve to fulfill the present requirements. Therefore, grounded in these concepts, the main goal of this paper is to introduce a teaching methodology via benchmark tests. Our methodology was conducted using the Hadoop framework, and it is innovative and proved effective. Our methods allow students to be exposed to complex data, system architecture, network infrastructure, trending technologies and algorithms. During the courses, students analyzed the performance of some computational architectures through benchmark tests on local and on the cloud. Along with this scenario, they evaluate the processing time of each architecture. As a result, our methodology proved to be a support learning method, which allows students to have contact with trending tools.