Probability Prediction of User Datagram Protocol (UDP) Upstream Throughput in a Network

Abstract

This work focuses on predicting the expected throughput value of the

network at a certain throughput range. A probability model for

multiple users’ real time data environment was developed for

estimating the probability of obtaining UDPupT on the network for

different signal ranges. Normal distribution was implemented to

obtain a general function for predicting the probability distribution

function (PDF) and cumulative distribution function (CDF) for

obtaining the different variables. The probability model provides an

additional way to predict a particular throughput value and not a

range of value, so we can describe the probability of obtaining

Signal to noise ratio (SNR) virtually in all the SNR cases is either

very high or low. The study shows that the probability model

developed from a combined multiple users’ data environment for

UDP protocol in a WLAN system trusted for estimating UDPupT has

a high throughput within the range of 8 to 9.99 Mbps.