Hello! Welcome to our fourth lesson.
In the last lesson, we explored the creation of ARPANET, the first large-scale packet-switched network. We saw how it was built for resource sharing and collaboration among researchers, and how it gave birth to foundational applications like email. However, ARPANET was just a single network. To get to the global network we know today, several crucial developments had to happen.
This lesson will guide you through those developments, addressing our learning outcome: Identify key milestones in internet history (TCP/IP standardization, DNS creation, commercialization). We will examine the three pivotal shifts that transformed a closed, government-funded project into the open, worldwide internet.
From One Network to a Network of Networks
The success of ARPANET proved the power of packet switching, and soon other networks began to appear, such as packet satellite and radio networks. This created a new challenge: these networks were designed differently and couldn't inherently talk to each other. It was like having multiple distinct train systems with different track gauges; a train from one system couldn't run on another's tracks.
A universal standard was needed to create a true "inter-net," a network of networks.
Milestone 1: TCP/IP Standardization
The solution to the internetworking problem came in the form of a new protocol suite: the Transmission Control Protocol/Internet Protocol (TCP/IP). Conceived by Vinton Cerf and Robert Kahn in 1974, it was designed to be an open standard that any network could adopt.
To understand the context and the brilliant design decisions behind TCP/IP, let's turn to a lecture from MIT.
23. A brief history of the Internet
This video, from a lecture by Prof. Hari Balakrishnan, explains the internetworking problem and introduces the fundamental design principles that made TCP/IP so successful.
Watch the segment from 24:38 to 33:10. Focus on: Robert Kahn's 'rules of interconnection' (e.g., each network is independent, best-effort communication). The role of 'gateways' (now called routers) in connecting networks. The decision to split the original TCP into two distinct protocols: TCP and IP. The 'end-to-end argument' as the justification for this split.
As the video explains, the design of TCP/IP was revolutionary. Let's break down the two core components:
-
IP (Internet Protocol): This protocol handles the addressing and routing of packets across the entire internet. Its job is to get a packet from a source computer to a destination computer, potentially crossing many different networks along the way. It operates on a "best-effort" basis, meaning it doesn't guarantee delivery. Think of it as the basic postal service: it knows how to read an address and get a letter to the right city and street, but it doesn't guarantee the letter won't get lost.
-
TCP (Transmission Control Protocol): This protocol works on top of IP to provide reliable, ordered, and error-checked delivery of a stream of data between two applications. If IP is the postal service, TCP is the certified mail service that tracks packages, ensures all the pages of your document arrive in the correct order, and asks for a signature upon receipt.
This separation of concerns is a cornerstone of network design. Not every application needs the reliability (and associated overhead) of TCP. For example, in a live video stream, it's better to drop a missing packet and move on rather than halt the stream to wait for a retransmission. Such applications can use the User Datagram Protocol (UDP) instead, which provides direct access to the basic "best-effort" service of IP.
To get a more detailed look at the principles behind TCP/IP, the following reading is an excellent resource.
A Brief History of the Internet
This foundational text from the Internet Society, co-authored by the creators of the internet including Vint Cerf and Bob Kahn, provides an authoritative account of TCP/IP's development.
Read the first six paragraphs, from the beginning of the article down to the end of the section listing the 'four ground rules'. This will reinforce what you saw in the video, particularly Kahn's critical design principles.
The adoption of TCP/IP wasn't immediate, but it gained unstoppable momentum. A key event was the "flag day" on January 1, 1983, when the ARPANET officially switched its core protocol from the older NCP to TCP/IP. This marked the true beginning of the modern internet.
Milestone 2: DNS Creation
With TCP/IP, any computer on the internet could reach any other computer using its unique numerical IP address (e.g., 18.62.0.96). This worked well for a few hundred computers, but as thousands of computers joined the network, it became unmanageable.
The initial solution was a simple text file, hosts.txt, that contained a list of all known computers and their IP addresses. Each institution maintained its own copy, and every night, computers would download the latest version from a central server. This was clearly not a scalable system.
The solution was the Domain Name System (DNS), developed by Paul Mockapetris in 1983.
This short video provides a very clear and concise explanation of the problem with hosts.txt and how DNS solved it.
Watch the segment from 14:56 to 17:11. Pay attention to the roles of Elizabeth Feinler in creating the naming registry and Paul Mockapetris in developing the automated, scalable DNS.
DNS is essentially the internet's phone book. It's a globally distributed, hierarchical database that translates human-friendly domain names (like www.mit.edu) into machine-readable IP addresses.
Its key features are:
- Hierarchical: Names are organized into a tree structure, with Top-Level Domains (TLDs) like
.com,.org, and.eduat the top. - Distributed: Instead of one giant
hosts.txtfile, the database is spread across countless DNS servers worldwide. No single server holds all the information. - Scalable: This distributed nature allows the system to grow almost infinitely without a central bottleneck.
The creation of DNS was a critical step in making the internet accessible and user-friendly, paving the way for its future growth.
Milestone 3: Commercialization
For most of its early life, the internet (first ARPANET, then its successor, the NSFNET backbone) was a US government-funded project. Its use was restricted by an "Acceptable Use Policy" (AUP) that largely prohibited commercial activities. So how did it become the commercial hub it is today?
The transition was gradual and intentional.
-
Stimulating Private Growth: The National Science Foundation (NSF) encouraged regional networks to seek commercial customers but prohibited that traffic from using the national NSFNET backbone. This created a business opportunity for private companies to build their own long-haul backbone networks to carry this commercial traffic.
-
Privatization of the Backbone: As these private networks (like UUNET and PSI) grew, the need for a government-funded backbone diminished. In 1995, the NSF decommissioned the NSFNET backbone, effectively handing over control of the internet's infrastructure to commercial providers.
The following text details this crucial policy-driven transition.
A Brief History of the Internet
Let's return to the 'A Brief History of the Internet' article to understand the specific policies that led to commercialization.
Read the section 'Transition to Widespread Infrastructure', focusing on the paragraphs that describe the NSFNET policies, the Acceptable Use Policy (AUP), and the eventual defunding of the NSFNET backbone.
This infrastructure shift coincided with the rise of consumer-facing Internet Service Providers (ISPs). Companies like AOL began selling internet access to the general public, typically via a dial-up modem connected to a phone line. This brought the internet out of universities and research labs and into people's homes, completing the transition to a public, commercial service.
Conclusion
In this lesson, we've charted the internet's evolution from a single experimental network to a global commercial platform.
Key Takeaways:
- TCP/IP Standardization: The creation of the open and robust TCP/IP protocol suite provided a universal language for different networks to communicate, forming a true "network of networks." The 1983 switch to TCP/IP is considered a key birthdate of the internet.
- DNS Creation: The Domain Name System replaced the unscalable
hosts.txtfile with a distributed, hierarchical system for translating human-readable names into IP addresses, making the internet vastly more user-friendly. - Commercialization: Through deliberate NSF policies that encouraged private competition, and culminating in the 1995 decommissioning of the NSFNET backbone, the internet transitioned from a government project to a commercially operated infrastructure accessible to the public.
Preview of the Next Lesson:
With the internet now established as a global, commercial network of networks, one final piece of the puzzle remains in our introductory module: the creation of its most popular application. In the next lesson, we will distinguish between the internet and the World Wide Web, exploring how the Web was built on top of the internet to make information universally accessible.