Kerberos: The Definitive Guide

The client-server model

The advent of the packet-switched computer network dramatically changed how users interfaced with computers. With the computer network, users all had an inexpensive computer on their desk, connected to all of the other computers in the organization through the network. Since the desktop computers still were not very powerful, specialized, more powerful server computers were used to provide services such as file storage, printing, and electronic mail to users.

This new computing model, widely referred to as client-server , provided some distinct advantages to both users and administrators (Figure 1-2). End users enjoyed increased usability since the desktop computers, while not very powerful, were powerful enough to present menu-driven interfaces to the resources of the server machines; a marked improvement over typing cryptic commands on a dumb terminal or line printer.

Figure 1-2. Client-server model

However, the faculty at MIT recognized that this new model required a dramatic new software architecture and a new way of thinking about computing: a way of thinking that recognized that computing power, rather than being centralized, was now distributed throughout the entire campus.

A major problem that the advent of personal computers and the network presented was that the end-user clients could no longer be trusted. In the traditional time-sharing model, end users ...