The growth and popularity of loosely-coupled distributed systems such as the
World Wide Web and the touting of Java-based systems as the solution to the
issues of software maintenance, flexibility, and security are changing the
research emphasis away from traditional single node operating system issues.
Apparently, the view is that traditional OS issues are either solved problems
or minor problems. By contrast, we believe that building such vast distributed
systems upon the fragile infrastructure provided by today's operating systems
is analogous to building castles on sand. In this paper we outline the
supporting arguments for these views and describe an OS design that supports
secure encapsulation of the foreign processes that will be increasingly
prevalent in tomorrow's distributed systems.