This is an installment of Ten Minute Tech; where I pick a technology related subject and then write a paragraph or two about what I know. I then pick and read a reference article related to the subject and then write another paragraph or two about what I learned. This edition’s topic is about the Advanced Research Projects Agency Network.

Before:

As well as I can tell it the Advanced Research Projects Agency Network (ARPANET) was the fist packet switched network and the direct predecessor of what we now know as the Internet. It was created sometime in the mid 1960’s by the United States’ Department of Defense.

My understanding is that it was the first electronic network of its kind because of its ability to communicate using packets of information. The idea of packets is that a single message or bit of information could be broken into several smaller chunks and then sent separately over the physical communication lines and then reassembled in one message once all packets were received. This could allow the simulation of simultaneous communication over a single line; understanding that a single line wouldn’t have to been tied up until the entire message was finished before accepting another one.

Article: Advanced Research Projects Agency Network

After:

In addition to what I already understand, the ideas behind the ARPANET were formulated in the early 1960’s by J. C. R. Licklider in his memos suggesting the idea off a “Galactic Network”. He was later given a position at ARPA and then started bringing others into the project. It wasn’t until 1968 that the full plan for the ARPANET was completed.

It was in 1968 that ARPA put out a Request for Quotation (RFQ); which was received as ‘outlandish’. Only twelve bids to complete the project were received and of which, only four were taken seriously.

The infrastructure was composed of (initially four) computers that were connected to dedicated, leased lines at 50 kbit/second. These computers (known as an IMP) handled all of the network routing and would sit at any location that wished to be connected and could support up to four machines.

The entire project including the hardware and route software was completed in nine months.

When the ARPANET was launched it was setup at four locations (potentially supporting connections for sixteen machines). The first of which being UCLA on October 29, 1969.

The IMP’s handled transmission by software called Network Control Program (NCP) which was designed to allow higher level software (protocols) to be written for it. It was this capability allowed for the creation of email and file transfer (FTP) protocols which we still use (a form of) today. Also voice transmission was created, but was never realized due to hardware difficulties of the time. It wasn’t until 1983 that TCP/IP replaced NCP, which is what we used today.

One common misconception about the roots of the network is that it was conceived out of a demand for an information network that could survive a nuclear disaster. However this was just a realized benefit after it was already up and running for a few years. It was truly conceived out of a lack of mainframe machines that researchers across the nation could utilize due to geographical restrictions.