Communications Group - Termly Report
====================   =============

Personnel

N. Millar has converted his half time for Computer Science to one day a 
week for this term and next.

Events of the last term
=======================

The large increase in network users this term has overloaded both Node 6 and
Node 9.  The overloading has two main causes, 1) The number of synchronous
lines supported by Node 6 has doubled in the past year with an increased
number of interactive terminals available. 2) Much of the increased traffic
was forced onto the Cambridge Ring through a gateway and a cpu dependant
ring interface.  During the term the network was restructured by a) bringing
a third FEP into service and b) moving the Ring FEP directly onto a Node.
This exercise caused a fair amount of disruption to the users, but
the network has been more resilient since the changes were made.  In 
addition to the hardware problems a few software problems have also been
fixed.

The Z80 software in the Ring DMA boards has been tested as far as it can be
with the existing PDP11 driver.  S. Hayes has offered to write the "proper"
PDP11 driver, hardware has been installed on the second filestore machine
and the software is being tested.  A couple of hardware items are required
before the boards can be used in service which will not be before the
5th Jan 1983.

The JNT PAD has been tested out in the development lab and when it is 
installed in the TP room it will be put into service with some "guinea pig"
users.

The FTP software has stabilised and is being heavily used; however the
addressing changes still have to be done.

The 11/60 gateway, Xgate, between the X25 switch and the Ring is working and
ready for the first real X25 connection.

The SERC Black box project is on target with the PDP11 software almost
complete.

The Coming Term
===============

      There is some apprehension that the network will still have serious
problems during the February peak and certainly will have to be significantly
upgraded before October '83.

The final decision on the topology of the X25 network has to be
reached and the appropriate hardware ordered.  The preparation of the X25
software will continue, although this is being hampered by the lack
of a development FEP.  The replacement of the NSI-RJE protocol, FTP,
has to be implemented in a workstation.

The ERCC Vax will be connected up to the X25 switch and will give some
operational experience of it and the gateway.

Changes to the TCPs to provide better support for video based programs
are being experimented with and negotiations for a long term Virtual
Terminal Protocol are proceeding in the academic community.

                                                     Brian Gilmore
                                                    December 1982