Uma abordagem baseada em políticas para contabilização e caracterização de uso global de grades computacionais

Abstract

Accounting and characterizing the use of computational grids constitutes a management task of paramount importance, especially when they are used in large scale,

involving many institutions and participants. Current grid computing management solutions are limited since they: (a) solely monitor the status of environment resources (like CPU and disk space consumption), omitting statistical and historical data about the execution of

applications; (b) do not offer support for gathering, processing and consolidation of information that is generated by heterogeneous grid technologies; and (c) do not allow the

specification of policies to distribute the collected information. To fullfil this gap, this work proposes an approach based on polices to account and characterize usage of grid computing

infrastructures, even when such grids are formed by heterogeneous middleware. The approach is materialized through an architecture based on the Web Services Distributed Management standard. This work presents the architec