Episode 10 Notes - American Networking
Aug 14, 2019
There isn't such a thing as a single event we can point to as the "birth of the internet", but we do get close to an event like that during the Cold War. That was when ARPANET, an early iteration of widespread networking, was designed and built. And while there were an uncounted number of contributors to the design of the final network, I think that the work done by Paul Baran at RAND is an interesting mark of the time that ARPANET came from.
Paul Baran was a computer scientist working on the problem of reliable networking at RAND in the early 1960s. The RAND Corporation is a civilian company that was started as an offshoot of the US Military with the express purpose of carrying out research for the Armed Forces. So it should come at no surprise that the work Baran was doing would have a militant streak to it. The "reliability" part of his research is much better termed as "survivability", that is investigating how to create a network that could survive an upcoming nuclear war with the Soviet Union. Luckily that war never came, but a lot of the choices that ended up forming into the modern internet came from this mindset.
In 1964 Baran would release his findings in a massive report, spanning multiple volumes. Inside he detailed a network very similar to the modern internet: a distributed net of computers sending data as binary packets. If you are familiar with networking, then that's basically a point for point description of todays internet. But what's interesting about Baran's work is the intent behind a lot of the choices he made. To underline this I want to quickly go over how Baran himself summarized his RAND report on networking. In the final volume he breaks his findings down to 7 bullet points:
"1) It appears theoretically possible to build large networks able to withstand heavy damage whether caused by unreliability of components or by enemy attack."
The "enemy attack" here is referenced throughout the paper, and it is made clear that the primary concern if for a nuclear-backed attack on American soil. But the fear of failing components is also relevant, you have to remember this is in an age not far removed from vacuum tubes. In both cases, the redundancy of Baran's proposed distributed network add a layer of safety and reliability.
"2) Highly reliable and error-free digital communication systems using noisy links and unreliable components can be built without exceeding the present-day state-of-the-art of electronic components--provided we use digital modulation."
It was never a given that any network would be digital, an analog network could have just as easily been proposed in the 50s or 60s. However, digital does have some key advantages specifically for long-distance communication. Since digital signals are effected much less by noise they can be carried by worse lines. In this case, the report is talking about using existing, and noisy, phone lines to carry digital data. Analog simply would have degraded too much over any distance on telephone cables.
"3) We are beginning to understand, or at least to appreciate, the cause of time delays and overloading phenomena in communication systems handling competing users with different levels of importance. There is a basis for hope that one day we may be able to automate highly sophisticated priority systems. Such systems may even be so effective as to provide the operational equivalent of exercised judgment."
Priority messaging isn't something that really exists in the internet today, but a lot of earlier networks had plans for this feature. Basically, this boils down to the fact that any network at this stage in development would be government funded, and any government wanted to be able to jam through priority messages for time-sensitive stuff(think nuclear launch codes).
"4) it appears that a proper direction in which to move in attacking the secrecy problem in large military and commercial communication systems, is to design the cryptographic provisions as an integral part of the digital switching system."
This part is interesting to me. Today we have security schemes like SSL/TLS that encrypt data as it's sent over a network, but those standards didn't come into being until the 1990s. But that just hides the data being sent, not where the data is going.
"5) Digital communication systems able to serve highly automated sources can be more readily designed from the viewpoint of bit-transportation systems rather than the conventional approach of a tandem connection of real-time links."
Digital data is better sent as binary chunks than as an unending stream. This is basically the idea of packet-switching, where a large message is broken down into small packets of data as it's sent and then reassembled. Since digital data can be represented as discrete numbers, you can do this easily and it gives you a lot of flexibility.
"6) One day in the future (and we are not foolhardy enough to predict an exact date), for economic reasons alone in the military environment it may be necessary to break away from existing analog signal communication network concepts in favor of all-digital networks."
We are solidly in the digital future, but I still don't think anyone can exactly put a date on when we fully migrated away from analog systems. It's wild for me to see a paper from the early 60s that is ostensibly about computing talking about the need to stop using analog systems. This feeds back into the point I was making earlier: it was never a given that the future would be digital.
"7) It is appropriate to redesign user input-output instruments, such as telephones and teletypewriters, for the described system in order to gain the full benefit that accrues to an all-digital communications network."
This may be my favorite part. Here, Baran is basically saying we will need all new technology to fully utilize an all new network. Instead of thinking about how a digital network could be made to work with existing technology, he is phrasing it as we need to change our current paradigm to work with a new digital network.
If you want to go deeper down the rabbit hole, the whole report is archived at the Internet Archive(https://web.archive.org/web/20101228070851/http://www.rand.org/about/history/baran-list.html) And if you want to hear more about the history of ARPANET, and how America first became networked, then check out my latest episode.
