Wireless Internet Network (WIN) and the Wireless Island Project

In the years since Nicholas Negroponte made his oft-quoted observation, it has become something of a cliché to maintain that that which is in the air should be in the ground and that which is in the ground should be in the air. It is, largely, a prophecy still waiting to be fulfilled.

Cheap medium band and cheap broadband remain elusive objectives. The actual supply side of internet services remains dominated by telephone companies which have a massive investment in an infrastructure which is inappropriate to deliver medium or broadband to end users – the so-called "last mile problem". Of course, it's only a last mile problem if you look at the network from the centre out. It is much more appropriate, we feel, to look from the user in. This makes it the first mile problem.

But for much of the world’s population, it’s not the first or the last mile which is the problem, but the failure over the last century to develop any kind of significant communications infrastructure.

Despite liberalization, the majority of telcos remain immured in a monopoly utility mindset: even where liberalization has occurred, progress towards implementing the implications of IP networking has been slow and costs have remained high. For example, reasonable medium band technology in the form of ISDN has been around for decades, yet in most cases (Germany is an exception) it remains priced beyond the means of small users – despite the well known economies of scale in value-added networks.

The overall situation is vaguely reminiscent of the attitude of the telegraph companies in the late 19^th century who used methods ranging from legal and legislative manipulation to violence to prevent the introduction of telephony services. In turn, most PTTs have, for over a century enjoyed the privilege of monopolies and the right to dictate to customers what both prices and services would be. It is unsurprising that most have been slow and clumsy in responding to the opportunities and challenges of packet switched networking.

In our view, the major market in volume terms for web services (beyond the obvious one of end users) is the small and medium-sized business sector. The web offers them the opportunity to access global markets at an extremely low opportunity cost. The overwhelming bulk of them currently access their web services through narrow band; while broadband does exist, in most areas it is priced well beyond the budgetary means of most smaller enterprises. Access to broadband would lead many to utilize centrally provided intranet/extranet/VPN facilities rather than develop their own. This in turn would spur demand for servers and workstations and networking tools: while most smaller firms are now to an extent computerized, it is largely with standalone computers or rudimentary networks. This, more than anything else, would spur on the development of a vigorous independent ISP sector. It is already known that in the US, where there is a wide range of choice, some 60% of businesses select an independent ISP rather than a telco as their web services provider, service being a key criterion.

Potentially, the commercial impact of the web in the Third World and former COMECON should be even greater permitting them to skip a generation of technology investment in copper networks.

In both scenarios, the key is a solution which can deliver medium and/or broadband at realistic prices, avoiding the inherent limitations of copper (or the lack thereof).

Several solutions are widely canvassed. DSL can provide medium band over twisted pair. Power lines have proved amenable (the ManWeb project in the UK), but are susceptible to their own "last mile" problem in North America. Cable modems are achieving wider distribution, but the cable companies are largely owned by media conglomerates with their own agendas. Similarly, satellite (usually limited to high speed down link) has been slow in implementation.

The authors of this paper both live in rural areas where digital telephony has only just been fully implemented. While they have had basic telephony for a century, full infrastructure development lags by decades in some cases. The same will be true of broadband. The same is true of large areas of the world.

There is an immediate demand for effective broadband networking solutions now. The immediate priority for most suppliers is high-density highly-developed areas where vicious battles will be waged by competing companies and technologies. This is not our focus. Our interest is in solutions for particular definable situations:

1. rural and small town North America
2. business concentrations, such as industrial parks
3. developing economies with a substantial and fairly sophisticated industrial/commercial base (more Egypt than Ethiopia; more Jordan than Chad) and a reasonable measure of market liberalization
4. industrial areas of Eastern Europe

1. a technical solution capable of delivering broadband networking at a realistic initial capital cost (in most of the situations we have identified a reasonable public sector contribution to initial infrastructural costs can normally be anticipated, usually from local/regional levels of government where decision making processes are often more informal and faster than from central government)
2. running costs which are acceptable amortized over reasonably small groups of customers yet commercially attractive to the operator: for example, an industrial park with 50 businesses paying an average of $200 a month = $10,000pcm = $120,000pa
3. a turnkey solution capable of management (with support) by the average level of competence one could expect in a local ISP
4. a package solution which delivers the key hardware software components for security, authentication, routing, caching, billing and communications
5. a vertically integrated network solution which makes it easy for end-customers to plug into the systems both at hardware and software levels
6. a channel solution which permits development of markets where it is traditionally fairly easy to sell but normally difficult to get paid
7. appropriate levels of technology. What most business customers need is the computer equivalents of pickup trucks, not Ferraris

We believe the most effective current solution is wireless networking. We propose to test this with a "working R&D project" which we call Wireless Island, for the reason that the proposed testbed is the Canadian Province of Prince Edward Island.

PEI is Canada’s smallest province with a population of about 130,000 in an area of under 3,000 square miles with a highest point of approximately 500 feet.

Internet penetration is high, largely due to effective government education and support programs that date back to BBS days.

As a part of the depressed Maritimes region, PEI is also the focus of several government investment and support programmes. Two key development agencies (the federal Atlantic Canada Opportunities Agency - ACOA - and the provincial Enterprise PEI) maintain IT specialists. The Atlantic Innovation Fund which followed from the Tomorrow's Wave consultation paper is attempting to address the infrastructural and economic divide which penalises Atlantic Canada.

We propose that the R&D not be limited – as is normally the case – to developing and testing the technical solution, but extended to cover the commercial and marketing solutions. To this end, we have incorporated a Canadian entity, Wireless Island Inc. It is our intention to create and run the Wireless Island project as a technical, commercial and marketing testbed on the theory that the best marketing proposition is a prototype which not only functions technically, but also commercially.

The technology exists now to provide integrated, secure, low cost solutions. We say solutions, rather than the solution quite deliberately. There are many different technologies and dozens of potential Beta/VHS struggles. Our solutions focused on the following criteria: cost, do-ability; focus on particular market niches (horizontal and vertical) and therefore shelf life; ease of implementation at both vendor and user level.

The product offering is an integrated turnkey wireless networking solution consisting of broadband web access relayed to users in either the 2.4 - 2.483 or the 5-6 gigahertz range, configured with the hardware and software to provide both secure web access and "standard" networking, initially offering users networking in the 2 - 12 mbps range. (The frequency ranges specified do not require licensing in Canada providing certain operating criteria are met.)

The end product would not be dissimilar to that of a turnkey ISP (and obviously the offering adapts easily to being a copper-based ISP), with the obvious distinctions that it is wireless and geared to secure networking, both IP and non-IP. The software mix continues to evolve in detail, but will include a mixed Linux and NT operating environment and a complete suite of content and network management tools.