COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	Dear Forest:
C00007 ENDMK
C⊗;
Dear Forest:

	This is a preliminary version of a proposal from the Stanford
University Computer Science Department requesting assistance from
D.E.C.'s External Research Program in the acquisition and development
of a file computer for the Department.  We believe that this project
will mutually benefit Stanford and D.E.C.
The general idea is the following:

	1. The Stanford Computer Science Department, like many research
and other institutions, operates a wide variety of computers all of
which need file storage.

	2. Present disk technology permits amounts of large enough
storage at reasonable prices so that one copy of any document prepared
by a person can be kept indefinitely and made available whenever wanted.
In particular, it is important to keep copies of all reports permanently
available.  Many files produced by machine are also small enough to be
kept.

	3. It is less expensive and more efficient to provide file
services on a specialized file computer.  This file system should
be able to grow with improvements in technology and in software for
keeping files.  Tape backup is still required and can best be done
on a large system.

	The system we plan would have the following characteristics:

	1. The file computer is a VAX 750 connected to the Ethernet.

	2. The disks are RA81s.  We propose to start with 4
gigabytes of storage and double it after a year.

	3. Two 6250 bit per inch tape drives are required for backup
and communication.

	4. The software we plan to develop will support file transfer
from the beginning and also work as a "leaf server", allowing a
program in any other machine to refer to a file without taking into
account whether it is on the machine's own disk or is on the file
server.  This will require software developed on our other machines,
and we anticipate that some machines will still be using explicit
file transfers when others are using the more or less invisible leaf
server.

	5. At some point the file server may perform some more
elaborate information storage and retrieval functions than just
the lookup of named files.

	6. At no time, however, will the file server have terminals
connected to it directly (except for its own system development) nor
will it do computation or users or user machines except that incidental
to storing and retrieving files.  Therefore, we expect to attain a
stable and reliable system in a short time.

	We believe that D.E.C. will eventually want to have such
a file system computer in its product line, and that collabortation
with Stanford will be mutually beneficial.

	The Principal Investigators of this project will be Professors
John McCarthy and Brian Reid.  We await your suggestions in putting
this proposal in more definite terms with respect to hardware and
financial relations.

	Sincerely,