Support for managing dynamically Hadoop clusters

Final master project developed in Barcelona Supercomputing Centre. This project focuses on the virtualization of Hadoop environments and the design of a piece of software for automatizing the configuration of the shared resources for Hadoop environments. In addition, this project uses a modified int...

Descripción completa

Detalles Bibliográficos
Autor: Nadal Bou, David de
Tipo de recurso: tesis de maestría
Fecha de publicación:2010
País:España
Institución:Universitat Politècnica de Catalunya (UPC)
Repositorio:UPCommons. Portal del coneixement obert de la UPC
Idioma:inglés
OAI Identifier:oai:upcommons.upc.edu:2099.1/9920
Acceso en línea:https://hdl.handle.net/2099.1/9920
Access Level:acceso abierto
Palabra clave:Cluster analysis
Hadoop
EMOTIVE
MapReduce
Virtualizatio
Anàlisi de conglomerats
Àrees temàtiques de la UPC::Informàtica::Programació
Descripción
Sumario:Final master project developed in Barcelona Supercomputing Centre. This project focuses on the virtualization of Hadoop environments and the design of a piece of software for automatizing the configuration of the shared resources for Hadoop environments. In addition, this project uses a modified internal Hadoop scheduler that adapts the available resources in the cluster for running jobs according to time restrictions. With the adapted internal scheduler and the virtualization capabilities the software developed in this project provides a flexible and self-adaptive service for Hadoop enviroments. In this report I introduce the context of the technology, the specifications of the system, the design of the solution integrated with EMOTIVE Cloud platform and the testing for showing the performance of the product.