Title:          Lessons from Project Athena

Speaker:        Jerome H. Saltzer
                Professor of Computer Science
                Massachusetts Institute of Technology*



                              ABSTRACT


Project Athena's primary goal is to improve undergraduate education
at M.I.T. with the application of networked engineering workstations.
To this end, it has deployed 850 such workstations and about 100
servers of various types in a networked client/server architecture.
Several technical developments were required in order to deploy this
system:

     - A window system that hides otherwise incoherent programming
       interfaces of bit-mapped displays (the X Window System),

     - a network authentication system that allows services to
       identify clients (Kerberos),

     - a network naming system that allows clients to locate services
       (Hesiod),

     - a storage model that integrates local disk, remote disk, and
       remote file system together with an unusual configuration of
       Berkeley UNIX that minimizes hand-tailoring of individual
       workstations (unnamed),

     - a system that locates users and distributes ephemeral notices
       (Zephyr),

and

     - a service management system that minimizes the effort of
       administration and coordination of the many servers and
       services (Moira).

As might be expected, quite a number of interesting problems were
encountered in creating this system: debates between proponents of
personal computers and proponents of engineering workstations,
management of deployment and operation at a large scale, preparing our
instituion to cope with a new technology, unrealistic expectations,
and working around the traditions and perspectives of the supplying
industries.



------------

*  From 1984 to 1988, Prof. Saltzer was Technical Director of Project
Athena.