Modelling and Simulation of Estimation Based Resource Allocation (EBRALLSR) in Programmable Active Networks
K. Vimala Devi* & K.M. Mehata"
Department of Computer Science and Engineering,
Krisnnankoil-626190, Tamil Nadu, India.
Chennai 600025, Tamil Nadu, India.
Active Networks allow customized computation on packets flowing through them. Resources in an active network mainly comprise of CPU and bandwidth. The inherent unpredictability of resources poses a significant challenge in providing quality of service (QoS) guarantees for data flows, which compete for processing resources in the network. An efficient allocation is required for the optimal utilization of the resources. In this paper, the resources are estimated using prediction models such as Single Exponential Smoothing (SES), and Adaptive Response Rate Single Exponential Smoothing (ARRSES). A comparison model of the estimated results is presented. An Estimation Based Resource Allocation algorithm (EBRALLSR) using the estimation model called Linear Least Square Regression (LLSR) is designed and the performance is measured for different applications with respect to delay measurements. The results show that EBRALLSR provides lower maximum delays to all application flows, when compared with other algorithms.
do so for
2000-2007 Professor F.R. Hall & Dr I. Oraifige, University of