This thesis is the result of a project whose objective has been to develop and deploy an
autonomous fault diagnosis system of software-defined networking (SDN) through Big Data
analytics tools and semantic web approaches such as Linked Data.
To do so, a system that recollects and process SDN data, as well as a diagnosis orches-
tration and visualization system both of the SDN data and the performed diagnoses.
The use case where our system has been deployed is based on a simulated simulated
SDN network environment and controlled by a SDN controller as Opendaylight. The developed prototype will collect data from this environment, it will perform a processing of
that network data, in such way we can detect symptoms of the network based on defined
parameters for the case study. Then, we will introduce symptoms in a diagnosis module by
phases, the first of them, to generate a initial set of hypothesis when a symptom is detected
and the second, to give a final conclusion collecting more information from the environment.
Furthermore, the prototype has a visualization system based on web technologies such
as Polymer or D3.js, that allows a network operator or a user a better network management
and its diagnosis.
As a result, we have a complete system be able to monitor and diagnose a telecommunications network based on SDN that can be interoperable with other applications or modules
thanks to the semantic approach given.