Some basic assumptions: As required by the question, let's suppose that the Internet didn't exist today. However, we assume that the computing and communications hardware have been developing at roughly the same rate as they have been in the past few years, and that hardware-wise, we are roughly at the same point we are today in the real world. We assume that the development in networking technology to date has been driven primarily by small-scale corporate intranets and by early research on the ARPANET [Cerf74].
Vision and Goals: Our company, Internet Inc., has the vision of a global Internet, a vast network of computers connected together using a standard set of protocols and the vision of a rich set of applications and services that provide value to end-users. The deliverable products of Internet, Inc. are:
Our goal is to be financially viable and, in the long run, profitable, quite unlike the design goals of the original ARPANET. Despite this fundamental difference in goals, our technical design shares a lot in common with the ARPANET, by leveraging its best technical features. The important ones are:
Market: Initially, we target corporate organizations and universities as users of this technology. Once this is successful, we plan to move to the residential market and target home users.
These are some important obstacles we need to overcome, and ways of overcoming them:
These are some key technical design/implementation decisions different from today's Internet and the essential strategic partnerships we need:
Technical design decisions:
Strategic partnerships:
[Cerf74] Cerf and Kahn, IEEE Trans. on Communications, vCOM-22n5, May 74, "A Protocol for Packet Network Interconnection", pp. 637-648.
[Horowitz94] M. Horowitz, T. Indermaur, and R. Gonzalez, "Low-Power Digital Design," Proceedings of the IEEE Symposium on Low Power Electronics, Oct. 1994, pp. 8-11.
[APM96] Intel Corp. and Microsoft Corp., Advanced Power Management (APM): BIOS Interface Specification, Technical Document, Feb. 1996.
[Burd96] Thomas D. Burd and Robert W. Brodersen; "Processor Design for Portable Systems" Journal of VLSI Signal Processing, 1996.
The obstacles to developing the Internet from scratch as a private corporation are numerous. The primary obstacles I anticipate would be:
I would address these obstacles in the following manner:
Building the infrastructure, tools, and applications - Building the Internet, obviously, is a monumentous task. It cannot be accomplished by a single company, or even a small consortia, but rather requires the ingenuity of an entire industry. I would try and develop an emergent technology wed based on my vision of the Internet. The web would include cable operators and telephone companies to provide access to the network, network companies to develop the infrastructure, server companies to host the web site, and software companies to develop the browsers and other required applications. I would own the network itself since it has the highest barriers to entry and would be most difficult to reproduce. I would enter into long term partnering agreements with the access providers and server sites to ensure that the link between the markets (servers) and the consumers is my network.
The primary difference in my network would be that it would be a far more symmetric network, with bandwidth allocated in such a way as to better support collaboration and real-time duplex interactions (audio and video). Such a network design will tremendously increase the value of the network and the benefits derived from it for both the consumer and commercial markets. Of course, such a network would require better protocols then TCP/IP. The network would need to better support audio and video to ensure that packets arrived in the right sequence and at constant intervals. Such a design would imply the ability to specify different service levels for different types of transmissions. The network would need to manage these different service levels and charge users based on the service level requested.
Financing and pricing - I would first turn to the government in search of some subsidies to help establish the infrastructure. While I would argue vehemently that the industry should not be regulated, I would look for subsidies based on the economic stimulus such an effort would provide the country. I would argue that the development of the Internet would provide the government with substantial revenue from tax dollars through the sale of products and services through this new distribution channel (obviously, I would be prepared to defend these projections with economic models, growth projections, etc. The fact I am pursuing this presumes there is economic benefit in it.) More significantly, the creation and maintenance of the infrastructure would generate millions of jobs (and the related income tax revenue) and be a significant boost to the entire economy (of course, I would present this to the President just before reelection as he is looking for a way to stimulate the lack luster economy.) Without some economic incentive from the government, such an infrastructure may never develop because of the enormous start-up costs required and the long lag in revenues.
Revenues from this model would come from a multi-level pricing model. In this scenario, a certain fee would be charged by server owners to have a site added to the networked server. In addition, I (the network) would charge a two part-tariff to the owner of each server connected to the network. The fee would be a base charge plus some incremental amount for each site hosted. This would encourage the server hosts to recruit more sites since it reduces their per site costs and thus improves their margins. I would provide free access on the consumer end initially until the network reaches 50% peak utilization and then I would introduce a fee that charges users a nominal fee based on sites accessed (not amount of data though, that is charged to the company sending the data to encourage more efficient sites). I would offer direct access to the network through some set of numbers as well as the option for value-added connection through ISPs (for faster access or additional content). This would help limit the price (especially the markup on my usage fee) ISPs could charge my customer base for access to the network. I would also charge the hosted companies (through the server companies) a fee for each user they send "data" to based on the amount of data.
Overcoming network externalities - The primary challenge will be overcoming huge network externalities. Therefore, the entire network will have to be in place before anyone gets any true value from it. I would overcome this in two ways. One, I would initially offer network access to the users for free. Second, I would provide some economic incentives for companies to host sites. These packages could include a period of free hosting, plus some money back guarantee if they are not satisfied with the return on their investment. I would also work with companies to develop marketing programs that truly provide value for consumers who use the network. We might develop rewards programs or special prizes for randomly chosen "network traffic". I expect this process to take along time.
Developing standards in support of the technology - In order to help reduce the impact of the network externalities and encourage the emergence of the network as a ubiquitous platform, I would also create a standards body to make it easy for others to tap into the network. Since the network will have the capability to charge on a usage basis, I want to encourage the use of the network for other activities such as real-time collaboration between companies. To broaden the markets for uses of the network beyond the WWW context of today,s Internet, standards will be needed to help developers more easily use the network to add value in organizations and for the public.