The Geography and Evolution of the European Academic Internet

The project aims to understand why certain academic nodes have grown out to major European hubs of high-end Internet and other academic nodes have not. We expect, apart from the traditional determinants of geographical location and population density, that three other factors affect the growth of bandwidth of a node. First, the mechanism of preferential attachment ? future connectivity is positively related to current connectivity? renders early entrants in the Internet backbone candidates for hubs. The time of entry in the Internet will thus affect its current bandwidth. Second, the more a host region is specialised in specific internet-intensive activities (both in academia and industry), the faster its bandwidth grows, because high-end Internet is not so much dependent on generic demand, but on high-tech applications in particular sectors. The project aims to model these relationships using graph theory, and to estimate the resulting model statistically using data on bandwidth, price (dependent variables), location, population density, year of entry, and specialisation in internet-intensive activities (independent variables). A third and final determinant of bandwidth growth of academic nodes concerns the institutional structure supporting hardware provision and Internet applications. The effect of institutions and policies is investigated for a number of case studies on successful and less successful nodes in The Netherlands. We expect nodes that involved stakeholders from academia within local policy networks to have been more successful in developing into an Internet hub, than nodes that relied on national institutions.