Fair QoS resource management and non-linear prediction of 3D rendering applications

Resource management in a grid has to able to guarantee commercial or industrial applications a personalized quality of service (QoS). To implement, however, an efficient resource allocation scheme, prediction of task workload is required. In this paper, we present an efficient algorithm for predicting the workload of 3D rendering tasks based on constructive neural network architecture. We also consider the QoS scheduling problem whose target is to determine when and on which resource a given job should be executed. We propose an algorithm for QoS scheduling, which allocates the resources in a fair way, and we compare it to other scheduling schemes such as the earliest deadline first and the first come first serve policies.