[ih] Historical fiction

[Vint Cerf]

Sytel,

BBN was the designer and operator of the ARPANET from the outset to
its termination in 1990.

I would distinguish the "subnet" of Interface Message Processors that
represented a major effort to design and implement a wide-area packet
network, from the hosts that used it for communication. The hosts were
programmed by staff at the universities and research institutions that
were first users of the network. So you had day to day operations of
the subnet by the "IMP Guys" at BBN and the daily operation of the
hosts on which applications were run. The motivation for building the
network arose from the vision of the power of non-numerical processing
expressed by Vannevar Bush, Douglas Engelbart, J.C.R. Licklider who
became the first director of ARPA's Information Processing Techniques
Office. Bob Taylor, then the director of the Information Processing
Techniques Office, was a strong advocate for the ARPANET and convinced
Charlie Herzfeld (then DARPA Director) to fund it and to bring Larry
Roberts from Lincoln Laboratory to ARPA to run the ARPANET development
program. Bolt Beranek and Newman was selected to design, build and
operate the network, drawing on the talents of Robert Kahn, Dave
Walden, William Crowther, Severo Ornstein, Frank Heart and many others
at BBN. The system was billed as a resource sharing network that would
allow all the computer science research teams, funded by ARPA, to
share computing resources and software. In truth it was an incredibly
successful experiment in packet networking and a major platform for
collaboration in computer science research. The host developments were
undertaken with the leadership of Stephen Crocker, then a graduate
student at UCLA working in Leonard Kleinrock's Network Measurement
Center. Steve and his colleagues were part of the Network Working
Group, formed and led by Steve. the NWG included dozens of
participants are the various institutions that participated in the
evolution and application of the ARPANET. For day to day operation,
you might consult with Alex McKenzie who is on this email list.

Eventually the ARPANET operation was transferred to the Defense
Communication Agency in 1975. Col. Heidi Heiden was a key program
manager in that time frame. During this period, the ARPANET served
both as an operational utility and also as a scaffolding that
contributed mightily to the development of the Internet.

vint

On Thu, May 10, 2012 at 11:15 AM, Sytel <sytel at shaw.ca> wrote:
> Hello again,
> Thanks for the go-ahead! Here's a rundown of the story concept and what we'd
> like to know...
>
> We've recently begun to write an open-ended piece of historical fiction
> about the history of the ARPANET - specifically its early days and the
> everyday/social aspects of living and working alongside such a remarkable
> network.
>
> Rather than viewing the ARPANET as a stepping-stone on the road to the
> modern Internet, we've been trying to look at the early network as an entity
> in its own right - not just a transitional phase, but as an important and
> ambitious project of its day, with great things in its future. To that end,
> we've taken the unique (we hope) approach of personifying the Arpanet
> (beginning life as a child, reaching adolescence by the 1972 demonstration),
> and having "him" interact with his developers and users, as he matures,
> increasingly fitting into the student society around him.
>
> In writing this, we've found ourselves in the awkward position of having to
> write about the members of the original Arpanet team. While we have
> attempted to go about this with the utmost respect, and doing as much
> research as possible, most of our sources are inadequate for this; they
> describe Arpanet only in the context of being something that would one day
> become the Internet. We're more interested in the very early social aspects
> and day-to-day operations of the lab, and how the Arpanet's future was
> envisioned in those days.
>
> We'd also like to be sure we're portraying everyone involved fairly
> and accurately, and we want to make sure we get the facts straight, as
> well as doing justice to the unique characters of the real people behind the
> story.
>
>
> These are some of the points we're wondering about; if you could answer any
> of these, that'd be great:
>
>
> How was Arpanet viewed/seen *at the time*? How did it fit in the context of
> the era? What was its future envisioned to be?

see summary above. Resource sharing! Experiment in packet switched
networking. I think it is fair to say that the "future" was captured
in the Internet concepts and pursued by NSF, DOE, NASA as well as
DARPA.  Many packet networks were developed in addition to the ARPANET
including X.25-based commercial systems, NSFNET, ESNET, NSINET, the
packet radio and packet satellite networks (of ARPA), Ethernets,
Proteon Rings, IBM Token Network, and many non-US academic and
research networks around the world. We were swimming in networks and
the Internet's TCP/IP protocols was one popular means of
interconnecting them all.

XEROX PARC deserves some of your attention because of its far-sighted
and highly innovative foray into high speed local networking (Ethernet
invented by Bob Metcalfe and David Boggs), personal computing, WYSIWYG
screen interactions, it's own variation of internetworking called PARC
Universal Packets (and associated host protocols), etc. The team at
PARC influenced the Internet designs, as did teams at the University
College London, Institute National Recherche d'informatique et
d'automatique. The National Physical Laboratory in Teddington,
England, influenced the ARPANET design (recommended very high speed
circuits to reduce latency) as well as inventing the term "packet"
(Donald W. Davies).


>
> What kinds of tests would be run on the Arpanet on a daily basis, i.e. what
> would be a typical day in the lab? And who would be in charge of performing
> what kind of test? What would happen in the lab other than tests?

The Network Measurement Center ran a wide range of artificial tests
(stochastic traffic generation) both to characterize the performance
of the subnet and to uncover ways in which the network algorithms
might fail. Bob Kahn had several theoretical scenarios of failure that
he tested and demonstrated at UCLA. Measurement data was used also to
validate queueuing models developed by Len Kleinrock or his students.
I ran many traffic tests with Bob Kahn and with many of Len
Kleinrock's students as the principal programmer at the Network
Measurement Center at UCLA.
>
> What were some of the first things it was actually used for other than
> testing the capabilities of the network? When did it begin to perform those
> duties?

Steve Crocker is a good source for the host level development story.
Larry Roberts pressed for usable results as the director of IPTO and
leader of the project. Larry asked Bob Kahn to organize a public
demonstration of the ARPANET for October 1972 as part of the
International Conference on Computer Communication in Washington DC.
This demonstration was a huge stimulus to getting things to work. Bob
Metcalfe organized and documented a wide range of demonstration
applications that the general public at the conference could try out.
Many of the engineers at BBN and graduate students at universities
funded by ARPA participated in the development and conduct of this
major demonstration. Seeing the list of applications would be useful
for you.

Ray Tomlinson announced his experiment in networked electronic mail in
1971; Larry Roberts wrote a TECO program to parse electronic maiil
called RD. Other email programs sprang up quickly. Standards were
developed and documented by David Crocker (Steve is his older
brother). MIT's Abhay Bhushan among others developed FTP. A messaging
development community emerged involving many others such as Eric
Allman (SENDMAIL), Dave Farber and his students, Einar Stefferud
(RIP), Jon Postel (RIP), John Vittal, ... well there are a LOT of
email pioneers to draw upon.

TELNET was the first protocol designed for remote access to computers
over the ARPANET and was a very important capability that allowed us
to use each others machines and software to work collaboratively.
Eventually we tested packet speech and packet video over the ARPANET
in the late 1970s and early 1980s.



>
> How did people react to its potential social aspects? Were they
> even considered at the time, or were they overshadowed by its potential as
> a scientific/knowledge-sharing system?

as soon as email came along, distribution lists were developed for
SF-Lovers (science fiction, not san francisco) and restaurant reviews
(Yum-Yum maililng list) and showed us all how quickly the system was
turned to social use.
>
> What did one of the old packets actually look like? We haven't been able to
> do much except extrapolate from modern packets.

there are detailed RFCs to be found in the Internet Engineering Task
Force records for formats, etc.
>
> What were some common problems that happened during the early days of
> testing? Other than the famous Login error.

routing and packet reassembly problems that led to lockups. IMP
failures that highlighted the need for checksums on packet content.
The need for larger and larger "windows" to allow many packets to be
"in flight" between pairs of correspondents on the network.
>
> Was the system ever shut down at night? At which point did they stop doing
> that, and what factored into that decision? At what point did the system
> become too important and decentralized to shut down *ever*?

It rapidly became essential for our work - I would argue that as of
early 1973 this was the case. It was not deliberately shut down.
>
> Had the ARPANET become personified in the manner described, what advice do
> you think you might have given it, at various stages of its development?

errr.
>
> It's been said that there are many fathers and mothers of the 'net. Are
> there any mothers that you might consider especially noteworthy in the early
> days?

Virginia Strazisar was the first to build a gateway for the Internet
linking, e.g. the packet radio network to the ARPANET.
Radia Perlman did extremely creative work on the formation of spanning
tree algorithms and on internet routing and recovery from
partitioning.
Karen Sollins at MIT worked closely with David Clark who became the
chief internet architect in 1982 (or maybe even 1981?).
Lixia Zhang from MIT and now UCLA did extraordinary work on network
design and architecture.
Deborah Estrin, also studying at MIT with David Clark, focused heavily
on applications of networks, especially mobile or mesh environments.

>
> What was the initial reaction to the first non-academic uses of the network?
> Any steps that were seen as wrong turns? Any that were seen as wrong turns
> at the time but worked out for the best eventually?

the first job advertisement caused a firestorm as did the first commercial spam.

>
>
> Please feel free to answer them in any order, or to leave out any that are
> too obvious, too time-consuming or just unanswerable.
>
> The story itself, for anyone interested, is here:
> http://forum.ultima-java.com/viewtopic.php?f=79&t=1353 It's a lot to read,
> though, and if you don't have much time, this chapter would probably be a
> good introduction to the characters and concept:
> http://forum.ultima-java.com/viewtopic.php?p=29299#p29299 Don't feel
> obligated to look at it, though; I'm not trying to hype it or anything.
>
>
> Thanking you in advance for your time,
> Sytel