摘要：At present, Internet emerges as a country’s predominantand viable data communication infrastructure. AutonomousSystem (AS) topology occupies the top position inthe Internet infrastructure hierarchy. AS resources are buildingblocks of this topology, and consist of AS numbers, IPv4 andIPv6 prefixes. Further, the resource requirement in each countryis dynamic and driven by various technical and socio-economicfactors. Hence, the organizational and national competitivenessfor socio economic development is reflected in AS growthpattern. Furthermore, to assess the competitiveness, futureexpansion, and policy development, there is a need for bothstudy and forecast AS growth. For Internet infrastructuredevelopment, understanding long term trends and stochasticvariation behaviour are essential to detecting significant eventsduring growth periods. In this work, we use time series basedapproximation for mathematical modeling, system identification,and forecasting to determine the annual AS growth.The AS data of five countries, namely India, China, Japan,South Korea, and Taiwan were extracted from the APNIC(Asia Pacific Network Information Centre) archive for thispurpose. The first two countries have larger economies andthe next three countries are advanced technological nations inthe APNIC region. The characterization of the time series isperformed by analyzing the trend and fluctuation componentof the data. The model identification is carried out by testingfor non stationarity and autocorrelation significance. ARIMA(Auto Regressive Integrated Moving Average) models withdifferent Auto Regressive (AR) and Moving Average (MA)parameters are identified for forecasting the AS growth ofeach country. Model validation, parameter estimation, pointforecast, and prediction intervals with 95 % confidence levelsfor the five countries are reported in the paper. The statisticalanalysis on long term trends and Change Point Detection (CPD)on Inter Annual Absolute Variations (IAAV) are presented.The significant level of change in variations, positive growthpercentage in IAAV, and higher percentage of advertised ASeswhen compared to other countries indicate India’s fast growthand wider global reachability of Internet infrastructure from2007 onwards. The correlation between AS IAAV change pointsand GDP (gross domestic product) growth periods indicates thatthe service sector industry growth is the driving force behindsignificant annual changes.
关键词：AS topology; statistical analysis; AS growth;forecasting; long term trend; inter annual absolute variation.
