Introducing Neutral Access Networks Alessandro Bogliolo Information Science and Technology Institute University of Urbino Urbino, Italy 61029 Email: alessandro.bogliolo@uniurb.it Abstract—Open access networks (OANs) have been recently proposed as a mean for bridging digital divide and enhance Internet penetration by enabling a fair competition among Internet service providers on a shared access infrastructure. This paper introduces the concept of neutral access networks (NANs), which are a special class of OANs conceived to grant positive externality to the shared infrastructure. Externality creates a positive feedback loop among users, service providers, and network operators which increases market penetration, motivates the development of new services, and promotes the deployment of new access networks. NANs are presented and discussed in terms of business models, market penetration, usage patterns, external benefits, technical feasibility, and legal issues. I. I NTRODUCTION Aggregate IP traffic is doubling every two years [1], Internet is becoming more and more integrated into our lives [2], and broadband connectivity is considered to be a top priority worldwide. Nevertheless, there is a stagnation in the broadband market due either to the lack of supply (incumbent operators are not motivated to support bandwidth-intensive innovations, while new entrants cannot afford the investments required to create their own access networks) [3], or to the lack of demand (stagnation in what users can do with networks ultimately leads to a stagnation in users’ demand for networks) [4]. Race, gender, education, purchase power, cultural background, and technical ability are additional determinants of broadband access [5], [6], that affect individuals’ perception of network potentials, the willingness to pay for bandwidth, and the ability to fully exploit it [7]. According to the Broadband Working Group of the MIT Communications Futures Program, there is a so-called broad- band incentive problem that comes from the inadequacy of the prevailing business model, which is based on vertical integration and on flat-fee pricing models [4]. Price is important both to understand the market and to shape it [5]. Flat rates were introduced in the narrow-band era to encourage penetration, but they induce user behaviors which are not correlated with the costs they impose to network operators. The diffusion of bandwidth-intensive applications and the extreme diversity of broadband traffic might induce operators to apply arbitrary restrictions on user behavior in order to monetize additional usage [4], ultimately violating neutrality and hampering innovation at the application level. Vertical integration, on the other hand, originated from the monolithic structure of early telephone networks [3]. Nowa- days it is still applied by most network operators in the attempt of realizing scope economies between physical access and online services [4]. Vertical integration, however, has several drawbacks: it limits innovation (by introducing dependences among different layers), it hampers competition (by raising entry barriers), it increases bandwidth costs (by reducing statis- tical resource sharing), and it endangers neutrality (by allowing incumbent operators to apply their own access policies) [3]. From a technical point of view, vertical integration fails in transposing to the business model the benefits of the layered nature of the network, based on TCP/IP and ISO/OSI models. Liberalization of telecommunications was not sufficient, per se, to create a true competition, because of the tremendous advantage of incumbent operators owning (or controlling) the local access infrastructures, which account for 80% of network costs in fixed deployments [8]. Regulations were then introduced in many countries in order to create the conditions for a fair competition by forcing the sharing of fixed access infrastructures (by means of local loop unbundling, line sharing, or bitstream access) [9]. In 2006 broadband penetration in countries with access sharing regulations was more than twice larger than in countries without [10]. The advent of wireless access technologies (WiFi, Hiperlan, WiMAX) and the allocation of significant public funding to address digital divide, is offering the opportunity for deploying new access networks. In order for such networks to make the difference, however, they have to be designed, deployed, and managed by avoiding vertical integration. The need for new business models able to separate network access from service provisioning and to enable a fair competition on a shared infrastructure is particularly apparent in case of access networks subsidized by public money and deployed with the main purpose of maximizing public utility [8], [11]. The concepts of operator-neutral networks (ONNs) [12] and open access networks (OANs) [3] were introduced to answer this need. OAN model encompasses a layered network architecture and a set of usage and trusting rules [3]. The OAN acts as an intermediator between users and services and it is composed of three main layers, represented in Figure 1: access islands (the local access infrastructures), operator-neutral backbone (the common infrastructure), and service providers (the entities providing online services, including Internet access) [8]. Battiti et al. introduced two sets of rules to be adopted in OANs in order to allow the network: i) to freely grow with needs and