-layer congestion control protocol such as TCP. The RED gateway has no bias against bursty traffic and avoids the global synchronization of many connections decreasing their window at the same time. Simulations of a TCP/IP network are used to illustrate the performance of RED gateways. lNTRODucTION I N high

This paper deals with a method using a specific class of neural networks whose learning phase is based on the Levenberg-Marquardt algorithm and which had been applied to the estimation of the traffic matrix (TM) of a large scale IP network. The neural network had been implemented with the help

Abstract—Despite a large body of literature and methods devoted to the Traffic Matrix estimation problem, the inference of traffic flows volume from aggregated data represents a key subject facing the evolution of next generation networks. This is a particular problem in large-scale carrier

, network routers can collect only relatively few statistics about the data. The central problem addressed here is to use the limited memory of routers to determine essential features of the network traffic stream. A particularly difficult and representative subproblem is to determine the top k categories

. Subspace methods such as Principal Component Analysis (PCA) can create accurate low-dimensional models. However, such models are not readily interpretable, as the principal components usually involve a large number of links in the traffic network. In contrast, the CUR matrix decomposition leads to low

. This paper presents a Neural Network Ensemble (NNE) for the prediction of TCP/IP traffic using a Time Series Forecasting (TSF) point of view. Several experiments were devised by considering real-world data from two large Internet Service Providers. In addition, different time scales (e.g. every five minutes

ii To my parents and grandparents who taught me to stand firm on human standards The study of Internet traffic behavior in a real IP network is the subject of this the-sis. Traffic Matrix of a telecommunication network represents the exchanged traffic volume between the source and destination nodes

global IP network measured over five-minute intervals. This allows us to do an accurate data analysis on the time-scale of typical link-load measurements and enables us to make a balanced evaluation of different traffic matrix estimation techniques. We describe the data collection infrastructure, present

neural network learning algorithm. By using the learning algorithm, a large data set can be split randomly and each slice data is handled by an independent neural network in parallel. The first experiment tests the algorithm’s learning ability on the benchmark of circle-in-the-square and compares

Abstract — In this paper, we present a practical VNT (virtual network topology) reconfiguration method for large-scale IP and optical networks with traffic matrix estimation considerations. We newly introduce a partial VNT reconfiguration algorithm with multiple transition stages. By dividing