It is widely assumed that, for reasons of efficiency, the various communication networks (Internet, telephone, TV, radio, ...) will merge into one ubiquitous, packet-switched network that carries all forms of communications. This view of the future is particularly prevalent among the Internet community, where it is assumed that packet-switched IP is the layer over which everything else will be carried. In this chapter, I present evidence so as to argue that this will not happen. This stance is controversial, and is difficult to make concrete, as any attempt to compare the various candidates for the transport infrastructure2.1 is fraught with lack of data and the difficulty of making apples-with-apples comparisons. Therefore, the evidence presented here is different from other chapters in this thesis. Observations, case studies, and anecdotal data (rather than controlled experiments, simulations and proofs) are used to take a stance and to predict how the network architecture will evolve.
Whatever the initial goals of the Internet, two main characteristics seem to account for its success: reachability and heterogeneity. IP, the packet-switching protocol that is the basis for the Internet, provides a simple, single, global address to reach every host, enables unfettered access between all hosts and adapts the topology to restore reachability when links and routers fail. IP hides heterogeneity in the sense that it provides a single, simple service abstraction that is largely independent of the physical links over which it runs. As a result, IP provides service to a huge variety of applications and operates over extremely diverse link technologies.
The growth and success of IP has given rise to some widely held assumptions amongst researchers, the networking industry and the public at large. One common assumption is that it is only a matter of time before IP becomes the sole global communication infrastructure, dwarfing, and eventually displacing, existing communication infrastructures such as telephone, cable and TV networks. IP is already universally used for data networking in wired networks (enterprise networks and the public Internet), and is being rapidly adopted for data communications in wireless and mobile networks. IP is also increasingly used for both local and long-distance voice communications, and it is technically feasible for packet-switched IP to replace SONET/SDH.
A related assumption is that IP routers (based on packet switching and datagram routing) will become the most important, or perhaps only, type of switching device inside the network. This is based on our collective belief that packet switching is inherently superior to circuit switching because of the efficiencies of statistical multiplexing and the ability of IP to route around failures. It is widely assumed that IP is simpler than circuit switching and should be more economical to deploy and manage. And with continued advances in the underlying technology, we will no doubt see faster and faster links and routers throughout the Internet infrastructure. It is also widely assumed that IP will become the common convergence layer for all communication infrastructures. All communication services will be built on top of IP technology. In addition to information retrieval, we will stream video and audio, place phone calls, hold video-conferences, teach classes, and perform surgery.
On the face of it, these assumptions are quite reasonable. Technically, IP is flexible enough to support all communication needs, from best-effort to real-time. With robust enough routers and routing protocols, and with extensions such as weighted fair queueing, it is possible to build a packet-switched, datagram network that can support any type of application, regardless of their requirements.
In spite of all the strengths of IP, this chapter will argue how it will be very hard for IP to displace existing networks. It will also conclude how many of the assumptions discussed above are not supported by reality, and do not stand up to close scrutiny.
The goal of this is to question the assumption that IP will be the network of the future. The conclusion is that if we started over - with a clean slate - it is not clear that we would argue for a universal, packet-switched IP network. In the future, more and more users and applications will demand predictability from the Internet, both in terms of the availability of service and the timely delivery of data. IP was not optimized to provide either, and so it seems unlikely to displace networks that already provide both. In this chapter, I take the position that while IP will be the network layer of choice for best-effort, non-mission critical and non-real-time data communications (such as information exchange and retrieval), it will live alongside other networks, such as circuit-switched networks, that are optimized for high revenue time-sensitive applications that demand timely delivery of data and guaranteed availability of service.
This is indeed a controversial position. Nevertheless, as researchers we need to be prepared to take a step back, to take a hard look at the pros and cons of IP, and its likely future. As a research and education community, we need to start thinking how IP will co-exist and co-operate with other networking technologies.
Section 2.2 provides a more detailed description of circuit switching and packet switching than in Chapter 1. It also describes part of the earlier work on these two switching techniques. Section 2.3 dissects some of the claims about IP, especially when compared to circuit-switched networks. This section tries to demystify those claims that do not hold up to scrutiny. Section 2.4 discusses the implications for the network architecture. Section 2.5 concludes this chapter.