Forecasting and Event Detection in Internet Resource Dynamics using Time Series Models

At present Internet has emerged as a country's predominant and viable data communication infrastructure. Autonomous System (AS) topology occupies the top position in the Internet infrastructure hierarchy. The AS resources which are building blocks of the topology such as AS numbers, IPv4 and IPv6 Prefixes are globally competitive because of the finite resource pool. The resource requirement of each country is dynamic and driven by various technical and socioeconomic factors. Hence the organizational and national competitiveness for socioeconomic development is reflected in the AS growth pattern. Furthermore to assess the competitiveness, plan for future expansions and to make policies there is a need to study and forecast the AS growth. As it is one of Internet infrastructure development indicators, understanding on long term trend and stochastic variation behaviour are essential to detect significant events during the growth period. In this work we use time series based approximation for mathematical modelling, system identification and forecast the yearly AS growth. The AS data of five countries namely India, China, Japan, South Korea and Taiwan are extracted from APNIC archive for this purpose. The first two countries are larger economies and the next three countries are advanced technology nations in the APNIC region. The characterization of the time series is performed by analyzing the trend and fluctuation component of the data. The model identification is carried out by testing for non stationarity and autocorrelation significance. ARIMA models with different Auto Regressive and Moving Average parameters are identified for forecasting the AS growth of each country. Model validation, parameter estimation, point forecast and prediction intervals with 95 % confidence levels for the five countries are reported in the paper. The statistical analysis on long term trends and change point detection on Inter Annual Absolute Variations (IAAV) are presented. The significant level change in variations, positive growth percentage in IAAV and higher percentage of advertised ASes when compared to other countries indicate India's fast growth and wider global reachability of Internet infrastructure from 2007 onwards. The correlation between AS IAAV change point and GDP growth period indicates that the service sector industry growth is the driving force behind significant yearly changes.