Implementación de una arquitectura tolerante a fallos en los servidores de producción de los sistemas de gestión académica de la Dirección de Desarrollo Académico (DDA) de la ESPOCH.

Abstract

The goal of this current degree work was the implementation of a fault tolerant architecture of the production servers in the academic management systems at the Academic Development Direction (DDA), such us: syllabus, analytic programs and planning where user management modules and roles were implanted which were developed using Scrum methodology obtaining a total of 17 user and 14 technical stories. Different tools were taken into account at the moment of creating the fault tolerant architecture which enabled to set up redundancy, stability and high availability, the main used tools are: Payara as application server whereby a cluster was built, and Apache as web server which together implemented a high availability using Pacemaker and Corosync as tools, moreover, a configuration file was created which works as load balancer, ultimately, a cluster was built through Citus extension which belongs to PostgreSQL, it has the main function of distributing database tables and do load balance of the requests which are sent to database manager. The final result of this research was the implementation of a fault tolerant architecture where the reliability corresponding to the standard ISO IEC 25010 was evaluated using Jmeter tool, with which 2 scenarios were made with 2 test cases which permitted to verify an error of 0,36 % of the total number of requests that cannot access to the system at the moment of processing 500 requests in normal conditions, whereas that at the moment of processing 2000 requests the error increases to 10,91%. To conclude, it is necessary to increase the hardware resources of the web server for serving more users. Finally, it is advisable to use scenarios for assessing a similar architecture.