This thesis proposal is about creating and evaluating a new resource
allocation method in large-scale computing environments (such as grids and clouds), by taking into consideration the complexity and particularities of the structures. For example, this could be done by means of analyzing the traffic generated at the grid sites, by investigating and modelling packet-level metrics.

Since computing grids normally reflect the panorama of a complex network, and the same can probably be said about cloud systems, we plan to apply complex networks concepts to improve the efficiency of the method.

We also intend to combine the aforementioned method with the JobPilot technique, currently used in the CMS/LHC grid system, and for which improvements are expected, both from the algorithm and the architecture viewpoints, as a part of the overall goal.

As an objective, this talk intends to present what we already did and what we plan to do for the future, which is surrounded by more general information on the context of my PhD.