How the Internet Came About
The Internet
requires different networks to be able to talk to each other, regardless of
which hardware, OS and programming languages they use. It’s only when you think
of it that way do you realize how difficult it must have been to achieve it!
Not just technically, but organizationally and politically.
In the 1960’s, the
US government (via ARPA) wanted inter-networking across universities. One, to
allow for ease of information sharing. Two, to enable sharing of computing
power! The second reason was also why many of the initial participants were
reluctant to join in, explains Walter Isaacson in Innovators:
“The universities in general did not want
to share their computers with anybody. They wanted to buy their own machines.”
When persuasion
failed, the coordinators switched to threats:
“There would be no more funding to buy
computers until they were hooked into the network.”
But the ARPA team
also looked to address legitimate concerns. They realized universities did not
want to “waste” their computing resources on routing logic (Where should the
message be sent? Which route should it take? etc). So ARPA decided to create
dedicated routers to make those decisions:
1)
This
solution took the load off the university computers;
2)
It
allowed ARPA to standardize the network;
3)
Routing
was now completely distributed, with no centralized hub.
The next question
was how were the messages to be sent? Via dedicated
lines (too expensive)? Store-and-forward
(waiting period would be unpredictable)? Packet
switching (break the message into smaller packets, each packet sent take a
route based on network congestion, and finally all the packets are reassembled
in the right order at the receiving end)?
They decided to go
with packet switching. It had the added advantage that it was decentralized.
But AT&T, the largest telecom company in the US, refused to buy into the
packet approach since, er… “Son, here’s how a telephone works”, one AT&T
executive said patronizingly. Eventually though, they agreed…
But packet
switching was still just an idea. Research teams were then selected to make it
a reality. Actually, it was left to humble graduate students!
“(They) kept meeting and sharing ideas
while they waited for some Powerful Official to descend upon them… But this was
a new age. The network was supposed to be distributed… Its inventions and rules
would be user-generated. The process would be open.”
That
open-to-anyone-to-comment/critique led to the now famous RFC (Request for
Comments) methodology: “friendly, not bossy, inclusive and collegial”. The
“grown-ups” never came, and the system just evolved by consensus to become what
we don’t even think about anymore: the bedrock on which all connectivity is
based.
Comments
Post a Comment