Improving Available Bandwidth Estimation using Averaging Filtering Techniques

End-to-end available bandwidth estimation is very important for bandwidth dependent applications, quality-of-service verification and traffic engineering. Although several techniques and tools have been developed in the past, producing precise estimations in real-time still remains challenging, mainly due to the lack of control on the measured network path and environmental conditions, such as noise and short term traffic burstiness. This paper presents VHF (Vertical Horizontal Filter), a modified Exponentially Weighted Moving Average (EWMA) technique borrowed from the financial world; our aim is to produce precise estimates of the available bandwidth, lowering the impact of noisy observations. The accuracy of this solution is compared to other EWMA techniques, showing that VHF has good performance and stability. Simulation results also show that VHF behavior is quite predictable and there is no need to fine tune the filter every time network conditions change.