Real-time Variable Bit Rate Video Traffic Using a Simple and Efficient Prediction Approach

The raising grandness of multimedia diligence has led to the increase in the usage of network resources. For assuring agreed Quality of Service (QoS) and for better network management, thorough analysis of traffic characteristics, modeling and prediction are crucial. In this process, a simple and effective traffic forecasting method which is suitable for time period prediction of MPEG encoded variable bit rate (VBR) video traffic. Our method is a variant of single exponential smoothing technique, which enhances prediction accuracy and in turn, enables better resource allocation and management. Results are compared with standard forecasting techniques like Single, Double exponential smoothing and Holt-winter’s. The proposed scheme attains relatively better prediction performance up-to 41%.