Modeling Autonomic QoS Control for Grid Service Using Petri Nets

The ability to offer and guarantee QoS to grid customers is of utmost importance for grid service providers. Due to the very uneven performance of grid resources over time and the unpredictable requirement of customers for grid service, the run-time QoS support is required to exploit a certain degree of self-management, or in other words to be partially autonomic. In this paper, a new approach, called autonomic QoS control, is studied. This approach introduces self-adjust configuration parameters into the grid system so that the QoS requirements of the grid services are constantly met. A general autonomic QoS control model is proposed, and its comprised components are expatiated. Three autonomic control scenarios are described and modeled using Controlled Stochastic Petri Nets (COSTPNs). Furthermore, a quantitative analysis for the performance of the model is given. We show that this solution can be helpful to design, analyze and develop autonomic QoS control system more efficiently.