Characterisation of wireless network traffic: Fractality and stationarity

Fractality and stationarity of a wireless network has been investigated in this paper by revealing the scaling pattern and nature of frequency fluctuation of the two vital parameters, the daily peak hour call arrival number and daily call drop number (03. 03.2004 to 31.10.2013), of a sub-urban local mobile switching centre, using some statistical methodologies. To have the knowledge about the fractality, Hurst Exponent for the series have been calculated using the methodologies like Visibility Graph Analysis, Higuchi's Fractal Dimension and General Hurst Estimation. It has been noticed that both the time series exhibit Short Range Dependent (SRD) anti-persistent behavior. Smoothed Pseudo Wigner-Ville Distribution (SPWVD) method has been used to unearth the stationarity/non-stationarity of the data-series where daily drop call time series shows evidence of non-stationary character while the busy hour call initiation series behave in a stationary manner.