Connection Era

The setting

Hardware: mainframes and minicomputers were expensive and scarce, held mainly by research institutions and the military; personal computers and workstations started to spread in the 1980s.
Networks: telephone lines and leased lines, with speeds measured in kbps; network protocols from different vendors and countries were mutually incompatible (such as X.25, DECnet, SNA).
Economics: against the backdrop of the Cold War, the US Department of Defense (ARPA) and the National Science Foundation (NSF) footed the bill, with essentially no commercial capital.
Users: scientists, engineers, and university faculty and students, numbering in the tens of thousands.
Policy: the NSFNET backbone was initially limited to research and education use; in 1992 US legislation opened it to commercial traffic, and in 1995 NSFNET was retired and the backbone was handed over to commercial operators; national research networks in Europe and Asia followed similar timelines.

The pressure

What needed to change

Expensive computing resources could not be shared remotely, so researchers had to go to the machine room and wait in line.
Different vendors and different networks could not interoperate, and every new connection required a dedicated conversion.
Network addresses relied on a single, hand-maintained HOSTS.TXT file, which became unmanageable once there were many machines.
Files were scattered across different machines; to find one you first had to know which directory on which machine it was in.
The network was built on the assumption that “everyone trusts everyone,” with no security design at all.

The building blocks

Packet switching and ARPANET → pain point 1: many computers sharing one line made remote login and file transfer possible.
Email (1971) → pain point 1: asynchronous communication between researchers, which later became the most widely used application on the network.
TCP/IP (1974 paper, standardized as RFC 791/793 in 1981, with ARPANET fully switching over in 1983) → pain point 2: a single “internetworking protocol” to connect heterogeneous networks, and the source of the word “internet.”
DNS, the Domain Name System (RFC 882/883 in 1983) → pain point 3: managing names in a layered, distributed way, replacing the manual hosts file.
The World Wide Web (proposed in 1989, made public in 1991): HTML + HTTP + URL → pain point 4: using hyperlinks to weave scattered documents into one web, so a single click jumps to a document on another machine.
The CERT/CC incident response center (founded after the 1988 Morris worm) → pain point 5: for the first time there was a dedicated organization for handling network security incidents.

The people

BackendBackend Engineer
Systems programmers, writing operating systems, network protocol stacks, and Unix tools.
OpsDevOps Engineer
System administrators, managing machine rooms, accounts, tape backups, and network connectivity.
SecuritySecurity Engineer
Incident responders, with the starting point being the Morris worm.

Trust & safety

GlobalThe US Computer Fraud and Abuse Act (CFAA) of 1986: enacted in 1984 with Pub. L. 98-473 and named the CFAA in 1986 by Pub. L. 99-474, now codified at 18 U.S.C. §1030 (current text: U.S. Code 2024 edition).
GlobalThe US Wiretap Act (1968) set an early statutory limit on intercepting electronic communications; the CFAA followed in 1986. Source

Selected moments

1969-10-29
ARPANET's first message is sent from UCLA to SRI
Sources: en.wikipedia.org
1971
Ray Tomlinson implements network email, using `@` to separate user and host
Sources: en.wikipedia.org
1974-05
Cerf and Kahn publish the TCP paper “A Protocol for Packet Network Intercommunication”
Sources: en.wikipedia.org
1981-09
RFC 791 (IP) and RFC 793 (TCP) are published
Sources: www.rfc-editor.org
1983-01-01
ARPANET fully switches over to TCP/IP
Sources: en.wikipedia.org
1983-11
RFC 882/883 propose DNS
Sources: www.rfc-editor.org
1986
NSFNET is built, becoming the backbone of the academic internet
Sources: en.wikipedia.org
1988-11-02
The Morris worm breaks out, and CERT/CC is founded shortly after
Sources: en.wikipedia.org
1989-03
Tim Berners-Lee submits the World Wide Web proposal at CERN
Sources: home.cern
1991-08-06
The World Wide Web project is made public
Sources: home.cern
1991-08-25
Linus Torvalds announces Linux
Sources: en.wikipedia.org
1993-04-30
CERN places the World Wide Web software in the public domain; the NCSA Mosaic browser is released the same year
Sources: home.cern

Looking forward

The network was connected and the Web existed, but it was still a tool for scientists, with no graphical interface and no commercial use allowed. In 1992 the US opened the backbone to commercial use, in 1993 graphical browsers appeared, and ordinary people and commercial companies began pouring in. The question shifted from “can I connect” to “can I understand it, and what business can I do with it.”

Dates and scope

Era boundaries

Boundary start: 1969-10-29 ARPANET sends its first message
Boundary end: 1993 CERN opens the web and Mosaic ships; commercial access begins in 1994

The setting

Hardware: mainframes and minicomputers were expensive and scarce, held mainly by research institutions and the military; personal computers and workstations started to spread in the 1980s.

Networks: telephone lines and leased lines, with speeds measured in kbps; network protocols from different vendors and countries were mutually incompatible (such as X.25, DECnet, SNA).

Economics: against the backdrop of the Cold War, the US Department of Defense (ARPA) and the National Science Foundation (NSF) footed the bill, with essentially no commercial capital.

Users: scientists, engineers, and university faculty and students, numbering in the tens of thousands.

Policy: the NSFNET backbone was initially limited to research and education use; in 1992 US legislation opened it to commercial traffic, and in 1995 NSFNET was retired and the backbone was handed over to commercial operators; national research networks in Europe and Asia followed similar timelines.

What needed to change

Expensive computing resources could not be shared remotely, so researchers had to go to the machine room and wait in line.

Different vendors and different networks could not interoperate, and every new connection required a dedicated conversion.

Network addresses relied on a single, hand-maintained HOSTS.TXT file, which became unmanageable once there were many machines.

Files were scattered across different machines; to find one you first had to know which directory on which machine it was in.

The network was built on the assumption that “everyone trusts everyone,” with no security design at all.

The building blocks

Packet switching and ARPANET → pain point 1: many computers sharing one line made remote login and file transfer possible.

Email (1971) → pain point 1: asynchronous communication between researchers, which later became the most widely used application on the network.

TCP/IP (1974 paper, standardized as RFC 791/793 in 1981, with ARPANET fully switching over in 1983) → pain point 2: a single “internetworking protocol” to connect heterogeneous networks, and the source of the word “internet.”

DNS, the Domain Name System (RFC 882/883 in 1983) → pain point 3: managing names in a layered, distributed way, replacing the manual hosts file.

The World Wide Web (proposed in 1989, made public in 1991): HTML + HTTP + URL → pain point 4: using hyperlinks to weave scattered documents into one web, so a single click jumps to a document on another machine.

The CERT/CC incident response center (founded after the 1988 Morris worm) → pain point 5: for the first time there was a dedicated organization for handling network security incidents.

The people

BackendBackend Engineer

Systems programmers, writing operating systems, network protocol stacks, and Unix tools.

OpsDevOps Engineer

System administrators, managing machine rooms, accounts, tape backups, and network connectivity.

SecuritySecurity Engineer

Incident responders, with the starting point being the Morris worm.

Trust & safety

GlobalThe US Computer Fraud and Abuse Act (CFAA) of 1986: enacted in 1984 with Pub. L. 98-473 and named the CFAA in 1986 by Pub. L. 99-474, now codified at 18 U.S.C. §1030 (current text: U.S. Code 2024 edition).

GlobalThe US Wiretap Act (1968) set an early statutory limit on intercepting electronic communications; the CFAA followed in 1986. Source

Selected moments

1969-10-29

ARPANET's first message is sent from UCLA to SRI

Sources: en.wikipedia.org

1971

Ray Tomlinson implements network email, using `@` to separate user and host

Sources: en.wikipedia.org

1974-05

Cerf and Kahn publish the TCP paper “A Protocol for Packet Network Intercommunication”

Sources: en.wikipedia.org

1981-09