Fernando de la Prieta is a member of BISITE group of University of Salamanca and has just defended his doctoral thesis.
Dr. D. Javier Bajo Pérez
Dra. Dña. Sara Rodríguez González
Cloud Computing, a known computing paradigm, has been emerging in recent years with great strength. This paradigm includes a new marketing model based on the pay per use revenue model, which has radically changed the business model on the internet and allowed companies and individual users to rent the computational resources they need at any given time. This new computational model has also allowed the production model for these computational resources to evolve very closely along the lines of the just-in-time production model, in which only the resources needed for the production of services are consumed according to the demand that exists at a given time. Within this context, the term elasticity is used in reference to the services being offered. For this to be possible, a vast number of underlying technologies have had to grow and develop in order to produce a technology niche with the ability to vary the resources associated with each service according to demand.
However, despite the undeniable advances that have been produced at a technological level, there is still much room for improvement in these systems. In this regard, the framework of this doctoral thesis proposes the use of multiagent systems, particularly those based on organizational models to control and monitor a Cloud Computing system. Due to this approach, one of the first in this field of research, new generation Cloud platforms will be able to include characteristics derived from Artificial Intelligence, such as autonomy, proactivity, and learning capabilities. To this end, a new model, unique in its conception, has been proposed; it can provide the organization with intelligent agents with self-adaptive abilities in execution time for open and highly dynamic environments in which there is also a high degree of uncertainty. With this model, the system is able to vary the computational resources associated with each service produced according to existing user demand, using the system’s own dynamic self-adaptation.