The post-PC era

Abstract

Many believe that the Post-PC is about new Information Appliances. We disagree. It is about a new kind of services-enabled Internet, able to deliver distributed, scalable, high performance and adaptable applications to diverse access networks and terminal devices.The focus of the Internet has been its protocols, primarily TCP/IP and HTTP. With the rise of the World Wide Web, and the explosion in the number of people with access to the Internet, the technical community has shifted towards new applications and network-based services to support them. We contend that while the protocol stack is well understood, the service architecture underlying these applications is evolving rapidly and somewhat chaotically. We review the evolution of the Internet, with a particular emphasis on the rapid developments of the last five years, and propose a candidate service architecture based on the emerging structure of the new Internet Industry..We divide the history of the Internet into two periods: before the rapid spread of the World Wide Web (WWW) and the emergence of graphical Web browsers, and afterwards. We set the pivotal year to be 1995, even though NCSA Mosaic appeared in 1993 and the WWW in 1991. The crucial event was the final privatization of NSFNet, enabling the formation of today's commercial Internet. This provided the critical flame that allowed commercial use of the Internet to ignite, brought on by the development of a real market: the simultaneous explosion in the number of individuals with access to the Internet (in part, due to lower cost PCs allowing higher home penetration rates) and the amount of increasingly useful content embedded in the Web.As late as 1994, the answer to “what is the Internet?” was “anything that runs TCP/IP protocols.” Its essential feature was that applications run on a small number of well engineered protocols and APIs, thus hiding the complexities of the underlying access networks. The Internet successfully spans technology based on twisted pair, coax cable, telephone, fiber optic, packet radio, and satellite access. This conceptual architecture was so essential to the Internet's success that it was dubbed the “narrow-waist” model.A raging debate developed in how best to converge data, television, and telephony networks to achieve a “National Information Infrastructure.” Rather unexpectedly, the killer application was not video-on-demand but information at your fingertips delivered via the WWW. The clear trend is to move telephony and video onto the Internet as particular applications, and to enhance the underlying networking technology to better support these. Virtually all of the long distance telephone carriers have already announced that they are migrating their networks to become IP-based.However, the most rapid innovation has been in the proliferation of a new service model on top of the Internet's narrow waist. This is being driven by the privatization of the Internet, leading to intense competition in access and backbone connectivity services (e.g., ISPs, NSPs, hosting, colocation, overlay networks, etc.), and the emergence of significant Web-based businesses: Portals, Content Delivery, E-Commerce, and in the near future, Entertainment.While much of the current development is being driven by consumer access to Web content and business-to-consumer commerce, there is also significant focus on using Internet-based services to integrate enterprise business processes. The Internet is rapidly becoming the means that allows a corporation to outsource many elements of its information technology operations, not simply to present a Web presence to consumers or business partners.Elements of this emerging service model include: network service providers (NSPs), Internet service providers (ISPs), Web hosting services, applications service providers (ASPs), applications infrastructure providers (AIPs), content delivery services, and so on. Some of these services are oriented towards corporate users (the enterprise market), others to ISPs, and still others to content developers.This presentation further elaborates on the service model we see emerging, within the context of the Internet as it exists in mid-2000. We also make predictions of how this may evolve from this point forward. Our focus is on the rich set of services above basic connectivity.The presentation is organized as follows. We set the context by giving a brief history of the Internet. Our particular focus is on the last five years, which has not been well documented in the technical literature. Next, we present the emerging Internet Service Architecture. We then use case studies to illustrate this architecture. Finally, we give some predictions of how this will evolve in the future.