ARPANET Technical Information: Packet Formats, Maps, etc
is to this century, and the next several, what Gutenberg's press
was to the last several: an invention which generated a major quantum change
in society - one of only a few such in human history (others being the
invention of writing, and of the telegraph/telephone).
Information technology is a 'force multiplier' for other all other
technologies, because it increases their effects many times over. In the
technology world, information technologies are to other technologies what
Gauss famously opined of mathematics - that it was the Queen of the sciences.
Technical information about the ARPANET can be difficult to locate online,
unless you know to search for magic strings like '1822' and 'NIC 8246'. There
is a very good, and quite accurate, book about the ARPANET, Where Wizards
Stay up Late, by Katie Hafner, but it lacks technical detail. There are
three very good technical papers which give a great deal of detail:
but those are not always available freely online via the original location of
- F.E. Heart, R.E. Kahn, S.M. Ornstein, W.R. Crowther, and D.C. Walden,
interface message processor for the ARPA computer network",
Proceedings AFIPS, 1970 SJCC, Vol. 36, pp. 551-567.
- J.M. McQuillan, W.R. Crowther, B.P. Cosell, D.C. Walden, and F.E. Heart,
in the Design and Performance of the ARPA Network",
Proceedings AFIPS, 1972 FJCC, Vol. 40, pp. 741-754.
- John M. McQuillan, David C. Walden,
ARPA Network Design Decisions",
in "Computer Networks", Vol. 1, No. 5, August 1977, pp. 243-289.
This site is an attempt to rectify this situation to some degree. It contains
some useful technical information (packet formats, maps, etc) along with
links to online technical resources. (It has been split up into multiple
pages since there are so many images.)
Note: The rest of this page assumes you know something about the
ARPANET. So if you don't know what an IMP is, you need to do some other
reading first to get some background.
BB&N produced maps of the ARPANET on a regular basis, as it was always
changing (usually growing, until its last years when it was slowly dismantled
and turned off). They were produced in both 'geographic' and 'logical' form.
Here are a substantial selection of both types (on separate pages, to reduce
the overall page size):
There is also
this interesting animated image,
which shows a sequence of early geographic maps.
In 1982, the DoD decided to split the ARPANET up into two networks: one,
called MILNET, would be an operational network for use by the military; the
other, to retain the name ARPANET, would be used by the research community.
The community was formally notified of this plan in December, 1982.
The plan was to group all the hosts on the ARPANET into two sets, and then
use the access control features in the IMP software to isolate each group
from the other; hosts in one group would only be able to send packets to
other hosts in their group. Access from one group to the other would happen
via 'mail bridges', routers which were 'attached' to both sets.
Once the network had been 'virtually' split into two networks in this manner,
work could begin on actually separating it physically - assigning each IMP to
one network or the other, and deleting all links which led from an 'ARPANET'
IMP to a 'MILNET' IMP.
The first phase of this split happened in October, 1983; the second phase was
not completed until
Prior to the split, in 1983, there were 113 IMPs in the ARPANET; after the
ARPANET/MILNET split, the MILNET consisted of 65 nodes, leaving the ARPANET
with 68 nodes.
For historical interest, a few sets of MILNET maps are included
There were two different kinds of packet formats used in the ARPANET: those
used for "packets" sent between IMPs (these are more like 'frames' in modern
parlance, since a single user "message" might be broken up into several
"packets" for transmission through the network), and those used for packets
(technically, "messages") sent between IMPs and their connected hosts.
There were two versions of the Host-IMP headers:
- The so-called "short leader" format, used from 1970 until roughly the end
of 1975 (I don't know the exact date, although I'm researching it - I have
the documents which give it), although it continued to be supported after
that date, but hosts using it could not reach all of the other hosts
in the network at some point shortly thereafter.
- The so-called "long leader" format, introduced around the end of 1975,
which allowed i) more IMPs in the network, ii) more hosts per IMP, and iii)
various other improvements.
You can find the packet formats laid out in detail
Online technical material
These are important original ARPANET technical documents, provided by major
institutions, and thus unlikely to suffer link-rot. (The amount of things
which one can find links to, but which '404' when you click on the links, is
amazing. Hopefully these will not suffer from that problem.)
Alas, the final major piece of NCP (NIC #7101, "Offical Initial Connection
Protocol") does not appear to be online at this point (although see
this early version,
which is almost identical to the final version).
In addition to these, the early RFC's (available ubiquitously across the
Internet) are also full of interesting technical material; RFC's 1 through
about 700 are almost entirely ARPANET related - only after about 700 does
Internet material start to appear. Most are quite short; the entire
collection (less the few long ones) can be quickly scanned in an evening.
(This page may eventually contain pointers to the most important ones.)
ARPANET Protocol Handbook
The ARPANET's entire
is described/specified in a thick volume entitled the 'ARPANET Protocol
Handbook'; it contains reprints of a number of documents, which together
fully give the protocols of the ARPANET. Its contents, along with links to
most of the documents in it, are given
Back to JNC's home page
© Copyright 2012, 2020 by J. Noel Chiappa
Last updated: 4/July/2020