Measurement and Prediction of Network Characteristics for Adaptive Video Servers

C. Spielvogel and L. Böszörményi (Austria)


Quality of Service (QoS), Network Measurements, Dis tributed Video Server


We present the extension of an adaptive distributed mul timedia streaming server architecture (ADMS) towards a Quality of Service based network infrastructure. The key component of the extension is the distributed active net work estimation and forecasting tool (DANEF). It is able to estimate, process and forecast bottleneck bandwidth, avail able bandwidth, delay, jitter and loss of a certain path be tween the streaming server and the requesting client. Us ing this information, the server node that offers the best QoS for a certain client is selected from a set of possible candidates. Active measurements are performed by send ing ICMP [15] packet trains and forecasts are performed by applying fast algorithms that need only small initialization sets. The accuracy of the measurements is achieved by ap plying an efficient and innovative filtering mechanism, the correctness of the forecasts is achieved by dynamically se lecting the best fitting forecast model and by considering the forecast error of previous samples. We have compared DANEF to other already existing measurement tools like BProbe, Cprobe, Pathload, Pathchar and Network Weather Service. As can be found in the evaluation part of this pa per DANEF is much better suited for estimating QoS char acteristics in the context of multimedia streaming than the other 5 tools it has been compared to.

Important Links:

Go Back