Thesis Overview Dynamic Scheduling on Grid Environments

The work carried out in the thesis describes efficient virtual environments management. One of the main contributions is a middleware scheduling optimization based on Grid environments [1]. Being the overall goal to get an optimal resource selection plus a coordinate task execution optimization. In particular, the research was related with the interaction between non trivial quality service and tasks distribution issues in meta-organizations. While linking allocation and policies belonging to virtual and local organizations. The idea was originated in the virtual and remote laboratories study for virtual spaces creation [2]. In many public and research institutes numerous computing resources are available, but these are not always accessible due to geographical distance, or do not have the ability to achieve a common goal. The concept of virtual space introduces an abstraction layer for managing these resources regardless of location and heterogeneity, thus making efficient use of available resources. During the thesis development, we had experienced and managed virtual spaces environment generation. An infrastructure has been defined, two types of laboratories were implemented, and a scheduling optimization for parallel applications has been proposed. Today these concepts have evolved, some of the ideas has been published and were implemented in commercial functional prototypes infrastructure, although research is still in the planning of data centers to thousands of computers