Shifting Paradigms with the Application Service Provider Model

I n the past four decades, several technological breakthroughs have made it feasible to sell computing as a service rather than a product. Supercomputers and clustering technologies have made huge amounts of raw computing power available, while time-sharing operating systems have made computing resources a divisible utility. Personal computers have educated generations of home and office computing users, and they now depend on such devices. Meanwhile, the Internet has become the world's largest data and computing-service delivery infrastructure, offering a new platform for network-centric computing. The World Wide Web has enabled the widespread growth of electronic commerce, while Web browsers have become universal graphical user interfaces for Internet-based services and thin clients. Component technologies have made it possible to produce huge numbers of reliable distributed software applications that benefit from specialization and greater economies of scale. Recently, application service providers (ASPs) have begun marketing the ASP model, which uses the Internet or other wide area networks to provide online application services on a rental basis—commercially delivering computing as a service. Figure 1 shows projections that indicate the ASP market will grow to more than $20 billion in 2003, yielding a compound annual growth in excess of 80 percent. 1 Yet for the ASP model to become the computing industry's mainstream paradigm , ASPs must make significant breakthroughs in networking infrastructure, computing technologies, and rental-based cost models and financial services. If successful, the ASP model, combined with the effect of distributed component technologies, will foster a new era of competitive network-based computing. Standardizing distributed components will make integration of applications with distributed components a common practice. Standardizing common application data formats or application programming interfaces will further break the ASPs' monopoly, thereby instilling new energy into the competitive network-centric computing platform. A networked economy will be possible that can easily integrate modular services from different providers. To date, the ASP model has been limited to industry applications, with the research community showing little interest in it. By surveying the relevant technologies that support ASPs, highlighting the technological challenges the ASP model introduces, and studying the impact of the ASP model on the general computing infrastructure, I hope to increase awareness of this important development.