perm filename ARPA.TXT[D,LES] blob
sn#124045 filedate 1974-10-09 generic text, type T, neo UTF8
∂9-OCT-74 1137 network site NIC
Date: 9 OCT 1974 1138-PDT
From: CERF at SRI-ARC
Subject: arpa progress report
To: les at SU-AI
cc: cerf
<CERF>ARPAPROGRESS.NLS;7, 9-OCT-74 10:53 VGC ;.D=off;PFit;.PSw=on;
ARPA Progress Report for Protocols Study 1
This report covers the period from 1 July 1973 to 30 June
1974. 1a
INTERNETWORK PROTOCOL DESIGN 1b
During this period, a design for an experimental
internetwork protocol was completed [1] and has been
circulated both to IFIP WG 6.1 and to other interested ARPA
research centers. In addition, an article describing the
basic concepts was published in May 1974 [2]. An updated
and more detailed design was prepared and circulated only
to the sites participating in ARPA sponsored
internetworking and is now undergoing further revision. 1b1
The participants in the internetworking experiment include
the University College London under the direction of Prof.
Peter Kirstein, Bolt Beranek and Newman under the direction
of Dr. Jerry Burchfiel, and the Stanford Digital Systems
Laboratory under the direction of Prof. V. Cerf. Plans were
laid to connect a TENEX system at BBN with a PDP-9 at UCL
and with a PDP-11 at SU-DSL, all running the proposed
Transmission Control Program [internetwork protocol].
Concurrently an experiment was outlined between the
National Physical Laboratory in England under the direction
of Dr. Donald Davies and the IRIA research center near
Paris under the direction of Mr. Louis Pouzin. In the
latter experiment, a Modula-1 computer at NPL is to be
connected to a CII 100-70 at IRIA running a protocol
proposed by H. Zimmerman and M. Elie of IRIA. 1b2
An agreement was reached regarding a common basic
addressing format for both protocols [3] and it is intended
that the results of these two experiments will be used to
settle on a final protocol which could be used to connect
all 5 sites. 1b3
In a concurrent effort, plans were made to study the
problem of connecting the TYMNET with the ARPANET using the
protocol proposed in [1]. During the period of this report,
only modest progresss has been made in this effort, but
enthusiasm for the project remained high. It is expected
that more concrete progres will be made during the second
year. 1b4
IFIP Working Group 6.1 met in June 1973 and the National
Computer Conference in New York, in September of 1973 in
Sussex as the NATO Conference on Computer Networks, and in
January 1974 at the Seventh Hawaii International Conference
on Systems Science. Plans were made to meet again at IFIP
1
74 in August 1974. WG 6.1 was reorganized into four
subcommittees to make working together easier: 1b5
Experiments Committee chaired by Prof. P. Kirstein 1b5a
Protocols Committee chaired by Mr. L. Pouzin 1b5b
Legal and Political Issues Committee chaired by Prof. F.
Kuo 1b5c
Social Issues Committee chaired by Dr. C. D. Shepard 1b5d
In another step to make W.G. 6.1 as self supporting as
possible, and in the wake of the reduced NIC services
offered by ARPA after 1 July 1974, all W.G. 6.1 members
were to pay for the cost of reproducing and mailing of
committee notes and reports. It was expected that this move
would also shrink the size of the group down to those who
were seriously interested in the work. 1b6
PDP-11 EXPANSION 1c
During January through March 1974, the PDP-11/20
installation was expanded using funds from the Joint
Services Electronics Program sponsored jointly by the ARmy,
Navy and Air Force. The PDP-11 facility now includes: 1c1
a) PDP-11/20 CPU with 28 K 16-bit words of memory
[maximum allowed] 1c1a
b) 5.6 M word Diablo 44 moving head dual platter disk.
One disk is removable; each will hold 2.8 M words. 1c1b
c) Unibus repeater to expand the number of Unibus slots
available 1c1c
d) Four asynchronous terminal interfaces, two for
hard-wired use and two for dial up modems. Two
Anderson-Jacobsen modems and two Direct Access
Arrangement telephone lines also installed. 1c1d
e) One OMRON microprogrammed CRT terminal with 4K byte
buffer memory. 1c1e
f) One card reader (not new) 1c1f
g) One upper case only printer (not new) 1c1g
h) Two Dectape drives (not new) 1c1h
i) One RS64 64K byte fixed head disk 1c1i
2
j) One 1024X1024 CRT (not new) with SU-DSL designed
controller and two joysticks (latter two are new) 1c1j
k) Three Texas Instruments Silent 700 portable terminals 1c1k
l) One 16 bit general purpose digital interface for
experimental device attachments 1c1l
m) One 50 Kbit/second modem with ARPANET VDH interface
for use with the ELF operating system [PDP-11 is
connected by VDH to SRI IMP] 1c1m
The ARPA contract pays for the rental of the Modems, TI
terminals, and maintenance on the PDP-11 during the summer
months; the Electrical Engineering Department of Stanford
University pays for maintenance during the rest of the
academic year. 1c2
SOFTWARE SYSTEMS 1d
ELF 1d1
In January, an ELF I system was installed. It proved to
be fairly reliable although it had a few bugs left. It
did not support the Diablo Disk or the dial-up
facilities. Nor did it have much of a File Transfer
Protocol [text files from the net could be printed on
the line printer]. The ELF system was used
intermittently during this period for access to the
ARPANET, but owing to shared use of the equipment for
academic projects, the ELF system was not up much of the
time. 1d1a
An attempt was made to integrate ELF with the Disk
Operating System (DOS), but this proved impossible since
DOS is configured for single user function and
simultaneous use of DOS with ELF caused ELF to lose
control of it critical interrupts. We investigated the
possibility of a Virtual Machine system, but the
PDP-11/20 does not have adequate hardware to support
virtual memory or privileged instruction trapping needed
for Virtual Machine Monitors. We concluded that only a
PDP-11/40 with hardware modifications similar to those
on the UCLA system would serve for such a Virtual
Machine system and gave up that approach as too costly
and time consuming. Consequently, the system still
alternated between DOS and ELF usage. 1d1b
File Transfer Protocol 1d2
During the summer of 1974, an FTP was written which
would accept MAIL files from the network and print them
on the line printer. The program was documented [4] and
3
plans were made to extend the system to full FTP
capability. 1d2a
Simple Minded File System 1d3
As an aid to the ELF user community, we proposed to
implement a simple minded file system which would permit
ELF to read or write contiguous files on the disk. The
detailed specification and implementation of this
package was seriously delayed owing to lack of
documentation of the new ELF II system to which SMFS was
to be interfaced. ELF II did not arrive during this
period, so only the basic SMFS design specification was
written using DOS I/O calls as the model for user level
interface. 1d3a
REFERENCES 1e
1. Cerf, V. G. and R. E. Kahn, "Towards Protocols for
Internetwork Communication" IFIP W.G. 6.1 Document #39,
September 1973 1e1
2. Cerf, V. G. and R. E. Kahn, "A Protocol for Packet
Network Intercommunication," IEEE Transactions on
Communication, Volume COM-22, No. 5,May 1974 1e2
3. Pouzin, L. [Revised V. Cerf], "A Packet Format
Proposal," IFIP W.G. 6.1 Document #48, January 1974 1e3
4. Haugland, T., "An Implementation of the ARPANET File
Transfer Protocol for ELF," Stanford University Digital
Systems Laboratory Technical Note #46, July 1974. 1e4
5. Cerf, V. and C. Sunshine, "Protocols and Gateways for
Interconnection of Packet Switching Networks" Proceedings
of the Seventh Hawaii International Conference on Systems
Science, Special Subconference on Computer Networking,
January 1974 1e5
6. Cerf, V. "An Assessment of Arpanet Protocols,"
Proceedings of the Second Jerusalem Conference on
Information Technology, July 1974 1e6
4
-------