Abstract

The emergence of the global Internet has dramatically broadened and changed the computing landscape. In particular, much of the value in contemporary computing systems derives from networked applications. Prominent examples include e-mail, Usenet news, the World Wide Web, and the many varieties of peer-to-peer networks.

However, the number of successful, large-scale, truly distributed, applications is exceedingly small. We argue that a major reason for this is that tools and other facilities available to aid the developers of these applications are inadequate. We propose a life-cycle for these applications, identify challenges that must be met to make the model viable, and detail our initial work towards meeting these challenges.