Research On Computer Network Bandwidth Measurement Technique

As numerous real-time applications are motivated continually, performance optimization and improvement of the real-time applications are required urgently, which depend on the estimation and allocation of the network bandwidth. Therefore, it is very important to measure the available bandwidth dynamically, fast, and accurately.The heterogeneity of the network links and the bursitiness of the cross traffic bring great challenge to the available bandwidth measurement. In this paper, an adaptive fast end-to-end available bandwidth (avail-bw) estimation algorithm (AABw) is developed. The proposed algorithm has the following characteristics: firstly, for various different cross traffic burstiness and load, the algorithm can dynamically tune the length of the probing packet train; Secondly, it improves the method of adjusting the rate of the probing packet train, not only it remains the benefit of the fast converge, but also it more suitable for the burst cross traffic situations; thirdly, we improve the algorithm of detecting the OWD trend. Through the simulation experiments, it shows that it can measure the avail-bw of the end-to-end path accurately and fast, suitable for the real-time applications.The avail-bw measurement based on the UDP probing needs to inject the extra traffic to the network. Inevitably, it is intrusive to the existing cross traffic. Besides, it must require the cooperation of the two ends, which significantly limits the usefulness, applicability, and the ease of deployment of tools. This works presents a novel avail-bw measurement algorithm, called SEAM, to which the bottleneck buffer estimation (AQI_BM) algorithm is embedded. The SEAM is based on the TCP data packets and runs on single end. Through the simulation experiments, it shows that it can measure the avail-bw accurately.At last, this paper presents the project of the avail-bw measurement based on both the AABw and the SEAM in detail.