In this work we deal with the issue of optimizing the global Quality of Service (QoS) of a Grid Federation by means of an aggregation model specifically designed for intelligent agents assisting Grid nodes. The proposed model relies on an algorithm, called FGF (Friendship and Group Formation), by which the nodes select their partners with the aim of maximizing the QoS they perceive when a computational task requires the collaboration of several Grid nodes. In the proposed solution, in order to assist the selection of the partners, a suitable trust model has been designed. Since jobs sent to Grid Federations hold complex requirements involving well defined resource sets, trust values are calculated for specific sets of resources. We also provide a theoretical foundation and some experiments to prove that, by means of the adoption of the FGF algorithm suitably supported by the proposed trust model, the Grid Capital (which respect the global QoS) of the Grid Federation is eventually improved.
A QoS-Aware, Trust-Based Aggregation Model for Grid Federations / Comi, A; Fotia, L; Messina, F; Rosaci, D; Sarne', G. - 8841:(2014), pp. 277-294. (Intervento presentato al convegno On the Move to Meaningful Internet Systems: OTM 2014 Conferences - 22nd International Conference on Cooperative Information Systems (CoopIS 2014) tenutosi a Amantea, Italy nel 27-31 October 2014) [10.1007/978-3-662-45563-0_16].
A QoS-Aware, Trust-Based Aggregation Model for Grid Federations
ROSACI D;SARNE' G
2014-01-01
Abstract
In this work we deal with the issue of optimizing the global Quality of Service (QoS) of a Grid Federation by means of an aggregation model specifically designed for intelligent agents assisting Grid nodes. The proposed model relies on an algorithm, called FGF (Friendship and Group Formation), by which the nodes select their partners with the aim of maximizing the QoS they perceive when a computational task requires the collaboration of several Grid nodes. In the proposed solution, in order to assist the selection of the partners, a suitable trust model has been designed. Since jobs sent to Grid Federations hold complex requirements involving well defined resource sets, trust values are calculated for specific sets of resources. We also provide a theoretical foundation and some experiments to prove that, by means of the adoption of the FGF algorithm suitably supported by the proposed trust model, the Grid Capital (which respect the global QoS) of the Grid Federation is eventually improved.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
