A Technical History of National Physical Laboratories (NPL) Network
Architecture - A Technical Tour
Motivation and Historical Background
NPL was originally founded in 1900 and a Mathematical
Division was formed with it at the end of World War II to research into
computing machinery and methods. In 1960, the group was renamed to the
Autonomics Division and was led by R. H. Tizzard. Early works of the division
included the ACE computer proposed by A. M. Turing and Duece computer
in 1946 and 1952, respectively. Donald Davies joined NPL in 1947 at the
young age of 23 and by 1960 was promoted to the position of Deputy Superintenedent
of the division. In 1961, a standard interface project was led by D. L.
A. Barber under Davies to "rationalize the connection of devices to computers
for process control and communication." This project is noted as the mark
for bringing NPL into researching in the telecommunications area. (Campbell-Kelly
222)
The concept of a 'store and forward' was inspired by time
sharing technologies and the 'inadequacy of data communications facilities
for interactive systems' in the 1960's. Donald Davies envisioned the possibility
of matching time sharing systems' performance with interactive system
if very short message sizes were matched the store and forward techniques.
In November of 1965, Davies wrote 'Remote on-line data processing and
its communication needs' that described this new concept for the first
time. It included ideas about how the existing telephony infrastructure
can support this digital data network. A longer document, 'Proposal for
development of a nation communication service for on-line data processing
was written in December that year to elaborate on the earlier paper. The
new paper included estimations and calculations on possible delay times
of the 'store and forward' system. Davies distributed this paper to a
small group of people who he felt would be interested in this type of
technology, but the response was not 'enthusiastic'. In the months to
follow, he gave a lecture at the National Physical Laboratory called 'The
future digital communication network' that described the design in more
detail. Davies gave a few more lectures in the subsequent years. It was
not until 1967 that he made his first open publication in the ACM Symposium
on Operating System Principles at Gatlinburg. And it was not until the
1968 IFIP Congress in Edinburg his idea was proposed to a larger audience
in a paper entitled 'Communication networks to serve rapid response computers.'
In less than 7 years after Davies' first described his packet switching
concept, a local network began operations. Finally in 1973, the network
was in full operations.(Davies)
The idea of packet switching implemented on a public data
network would help to resolve the issue of incompatibilities between computers
sharing information.(Davies)Davies also noticed
that the telegraph system's message switching techniques were anaglous
to the communication networks. But instead of transmitting an an entire
message, shorter broken-up 'packets' can be sent to eliminate the delay
faced by telegraph networks.
