The fine folks at Merriam-Webster tell us that one of the definitions of “protocol” is (3a): “a code prescribing strict adherence to correct etiquette and precedence (as in diplomatic exchange and in the military services)”
Showing my age – the first time I heard the word “protocol” was back in the last millennium when a little movie (“Star Wars” 1977) had a throw away line about a character being a “protocol droid” – which kind of illustrates the point I’m working towards
The English language has a LOT of words (for various reasons). Those words also tend to change in usage over time.
e.g. Henry Kissinger was (probably) and expert on “protocol” back in the last half of the 20th century. That would have been “diplomatic relations at a nation state level” – but you get the point
Computers
While I’m at it – the English word “computer” USED to refer to a human being that performed complex calculations. Back in the first half of the 20th Century you might have heard folks talking about a new “electronic computer” to differentiate from “human computers.” (fwiw: I recommend “Top Secret Rosies: The Female ‘Computers’ of WWII” to technology and history enthusiasts)
In the early days of “electronic computers” – the “computer” tended to be very large – e.g. ENIAC weighed 30 tons.
There was obviously no need to “network” computers together because, well, there simply weren’t that many “computers.”
Leap forward to the 1970’s and transistors had allowed the size of “computers” to decrease (e.g. an IBM 360 model 65 only weighed between 4,290–8,830 lbs), and Ethernet had been invented (1973).
Depending on your bias just HOW import Xerox PARC was to “modern networking” is debatable – but there is no questioning that it WAS important.
From a “techie” point of view you can’t have a “network” until you have the means to connect endpoints together. e.g. the POTS (“Plain Old Telephone Service”) functioned by analog transmission over copper wire. Someone had to run (and maintain) all of that wire for POTS to function.
Of course the folks developing Ethernet weren’t thinking about “networking computers” – they wanted to network Xerox copiers together.
“Way back then” (the 1960s/70s) the only folks with “computers” were LARGE corporations, military installations, and academic institutions. Ethernet allowed ARPANet (i.e. the origin of “the internet”) to function
Terminals
WHY would anyone want to create a “network?”
With POTS you might say that the goal was “fast communication” – i.e. picking up the phone and dialing a phone number gave you access to anyone on the network, which is much faster than writing a letter, mailing the letter, waiting for a response, etc.
In the old “mainframe only” days remote access to the system might be accomplished with “dumb terminals” – but that wasn’t really a “network”, more of an “extension.”
A “dumb terminal” was “dumb” because it couldn’t do anything useful disconnected from the mainframe. It might have looked like an independent device – with a keyboard and screen, but it was connected directly to the mainframe and had no “computing” capability by itself.
Under the “nothing new under the sun” concept – various flavors of “low powered devices” connecting to resources over the Internet has brought back the concept of “dumb terminals” – but that isn’t important at the moment …
That “mainframe” probably had a large and noisy “line printer” attached. This also wasn’t “networking” – just a device attached to the main computer.
Just for fun – hunt up any movie set in a 1970’s “newspaper office” and look for the combination of “typewriters” and “terminals” on reporter desks.
(e.g. “Superman” 1978 has a brief scene in the “Daily Planet”)
Those “terminals” might have had the ability to hold a “text document” in memory for editing, but not much else. The “terminal system” would have been an improvement on using a typewriter, but wasn’t a “modern word processor.”
Old newsrooms would have had “copyboys” waiting around – the new “computer system” might have let reporters print off their story remotely (removing a primary job of “copyboys”) but again, that wasn’t “networking”
Local Area Networks
Jump to the middle 1980’s and the first “personal computer networks” were centered around sharing a printers and files.
If computers are going to “talk to each other” and share resources then they need to “talk the same language” – i.e. they need a “protocol”
Once upon a time it used to be said that “No one ever got fired for buying IBM.” Meaning that selecting IBM as your vendor may not have been the lowest price – but IBM was a proven entity with a long history of quality performance in the marketplace – so IBM was “trusted.”
Which is important to this story because the “corporate world” didn’t REALLY accept “personal computers” until IBM came out with the IBM PC in 1981.
As the name implies the “PERSONAL” computer was a stand alone device.
If you wanted to print a document on a remote printer or share a file “sneaker net” was probably the most popular solution in the early 1980s – which meant “copy the file to a disk and walk that disk to the other device.”
We need to back up a little bit and point out that an up and coming software tools company named “Microsoft” provided IBM with an operating system for the IBM PC.
Microsoft, Inc bought an existing Operating System (OS), did some work on it, and agreed to license the “Disk Operating System” (D.O.S.) to IBM. Which is another story for another post specifically about OSs …
The important point is that D.O.S. had no “networking” capability. If you wanted to create a “local area network” of personal computers running D.O.S. you needed a “Network Operating System”.
For our purpose today we can define a “local area network” (LAN) based on geography. The difference the LAN and the old mainframe/dumb terminal system is that the devices on the LAN are independent/have stand alone capability.
The “old system” didn’t just disappear. As a “techie” in the late 1980s I remember putting “terminal emulator” cards into brand new personal computers – which allowed them to function as dumb terminals (as well as use Lotus 1-2-3 or WordPerfect to name drop old software).
IBM released a proprietary “local area network” protocol called NetBIOS in 1983, the same year the first version of “NetWare” was released.
Novell NetWare came to dominate the LAN marketplace in the late 1980s and early 1990s. NetWare used a proprietary protocol called IPX/SPX.
Microsoft disrupted the “LAN” marketplace in 1992 when they released the first functional version of “Windows” (v 3.1). For the record, Microsoft Windows 3.11 (Windows for Workgroups) used an updated version of IBM’s NetBIOS (i.e. NetBEUI was the “NetBIOS Enhanced User Interface”).
NetWare had a large installed base of “LAN”s – so Microsoft including basic “local area networking” in Windows 3.1 wasn’t IMMEDIATELY a problem for Novell.
The lurking unknowable “black swan” networking protocol in the early 1990s was TCP/IP – which was that “open source” protocol that the “Internet” thing was using. The “Internet” at that time consisted almost entirely of college campuses and military research institutes.
How “TCP/IP” became synonymous with “network protocol” is another story …
“File and print” sharing is still the “bread and butter” of BUSINESS personal computer networks. The average HOME network is (just me guessing) probably using the LAN to share an Internet connection – but again, that is me guessing based on personal observations.
Routing
Imagine that you work for “Big Company, Inc” which has offices in New York City, Chicago, and San Francisco. Each office has a LAN but also has to access resources at the other two locations.
If you communicated with the good ol’ written word on pieces of dead trees – then each office for “Big Company, Inc” might have a “mail” room were ALL correspondence gets sent.
So if Alice in the Chicago marketing department sends a note to Bob in the Chicago Human Resources department – the note would be sent to the “mail room” where someone would look at the address, and then send it to Bob on the local network.
If Bob in Chicago need to check with Tim in New York about Alice’s note – then Bob (in Chicago) sends his new message to the “Chicago mail room” – the folks the “Chicago mail room” look at the “to” address, see that it isn’t in Chicago and then put the message on its way to the “not local” network.
Eventually Bob’s message gets to the “mail room” in New York, and they forward it to Tim. Then that whole process would work in reverse when Tim sends his reply to Bob.
That was (kind of) the way “snail mail” used to work. At a smaller and faster level – this is also how “packet switched” computer networking functions.
Of course the “computer network protocol” has to have some way of “routing” traffic for “non LAN” destinations.
The limitation for NetBEUI was that it didn’t allow for “routing.” IPX/SPX was “routable” but was also “proprietary” (as in “not free”). TCP/IP is routable AND open source – so …
