Automatización del aprovisionamiento de servidores y despliegue de servicios en un centro de datos

Abstract

The aim of this project is to provide a solution to two scalability issues that arise in big computation infraestructures: Inefficient resources utilization in the data center and the increase of the management complexity. Regarding fleets of servers, you don't have to go back far in time to verify that both configuration management and resources allocation polices have evolved very fast. The server is no longer the gravitation center and the data center consolidates as the new abstraction layer where computation happens. This project presents two platforms in simbiosis, the bootstrapping platform and the exploitation platform. The former is used to bootstrapp and manage the latter, which in turn unifies the data center resources and performs dynamic resources allocation based on demand. The project has a balanced effort of design, development and integration phases. The exploitation platform is larger in terms of number of components however, it just required an integration effort. All the building blocks are detailed in the project as well as its functions and its interactions. Despite being the smaller one, the bootstrapping platform requires the higher design and development efforts. All the building blocks are detailed in the project as well as its functions, building recips and full structural and procedural diagrams. The project ends with an installation guide that helps the reader to deploy a bootstrapping platform which in turn can be used to deploy an exploitation platform.