3. The Internet and the World Wide Web: Origins and Architecture

To understand what Web3 claims to be — and to judge those claims fairly — one must first understand the thing it proposes to revise. Web3 is, by its own account, an attempt to re-architect the web. Yet "the web" is routinely confused with "the Internet," and the design philosophy that animated both is often mythologized rather than examined. This chapter establishes two foundations the rest of the book will lean on: first, the precise technical distinction between the Internet (a network of networks) and the World Wide Web (one application running on top of it); and second, the historical fact — frequently invoked, sometimes exaggerated — that both systems were born from a decentralized, permissionless design philosophy. That second point matters thematically, because Web3's advocates present their project not as an invention but as a restoration: a return to origins from which, they argue, the web has strayed. Whether that narrative holds up is a question for later chapters. Here, the task is to get the history and the architecture right.

The Internet and the Web are not the same thing

In everyday speech the words "Internet" and "web" are used interchangeably, and for most purposes the confusion is harmless. For this book it is not, because the entire Web3 argument turns on a distinction between the underlying network and the applications built upon it.

The Internet is the network of networks — the global system of physical cables, wireless links, routers, and shared rules that allows any connected computer to exchange data with any other. The World Wide Web is one application that runs on that network — a system of interlinked documents and resources, retrieved through web browsers, that happens to be the Internet's most famous use.

A useful analogy is a city's road system versus the activities that use it. The Internet is the roads: the asphalt, the intersections, the traffic rules that let any vehicle reach any address. The Web is one kind of traffic on those roads — say, the postal service delivering letters to street addresses. Email is a different kind of traffic; so are video calls, file transfers, and the messaging inside a smartphone app. All of these ride on the same roads, but none of them is the road system. The Web is to the Internet roughly what the postal service is to the road network: an immensely important service that uses the infrastructure but is not identical to it.

In plain terms: the Internet is the plumbing; the Web is one of the things that flows through the pipes. When you open a banking app, stream a film, or send an email that never touches a browser, you are using the Internet without using the Web. When you visit a page in a browser, you are using the Web, which is in turn using the Internet.

The analogy breaks down in one respect worth flagging: roads are owned and maintained by governments, whereas the Internet has no single owner. It is a voluntary interconnection of tens of thousands of independently operated networks that agree to speak a common language. That ownerless quality is not incidental; it is the historical and architectural heart of this chapter, and the feature Web3 most insistently invokes.

The chronology reinforces the distinction. The Internet's foundational protocols were operational by 1983 (ETHW, n.d.). The Web did not exist until a proposal written in 1989 and a public launch in 1991 (Berners-Lee, 1989; CERN, n.d.). For roughly two decades, in other words, there was an Internet but no Web. The Web is the younger, narrower thing — a layer of software added long after the network beneath it had matured.

ARPANET and the idea of packet switching

The Internet's lineage runs back to ARPANET, an experimental network funded by the United States Defense Department's Advanced Research Projects Agency (ARPA, later DARPA) in the late 1960s (Britannica, n.d.). ARPANET was built to let a handful of expensive, incompatible research computers at universities and laboratories share resources over telephone lines. It was a research instrument, not a consumer product, and certainly not a commercial one.

Its decisive technical innovation was packet switching.

Packet switching is a method of moving data across a network by chopping each message into small, individually addressed chunks called packets, sending them independently across whatever paths are available, and reassembling them at the destination.

Contrast this with the older technology it displaced, circuit switching, the principle behind the traditional telephone system. In a circuit-switched call, the network reserves a single continuous electrical path between two parties for the entire duration of the conversation; that path is dedicated and idle whenever no one is talking. Packet switching abolishes the dedicated path. A useful analogy: instead of chartering an entire private train to carry one letter from one city to another and back, you write the letter on many postcards, drop them into the ordinary mail, and let each postcard find its own way through the system — some by one route, some by another — to be put back in order on arrival. Each postcard carries the destination address, so the network needs only to forward each one toward where it is going; it does not need to understand the whole message or hold a route open.

This design has two consequences that echo through the rest of this book. First, it is efficient: many conversations can share the same lines, because capacity is used only when packets are actually moving. Second, and more importantly for the Web3 narrative, it is resilient and decentralized by construction. Because each packet is routed independently and can take any available path, there is no single point whose failure stops the flow. If a link or a node goes down, packets simply route around it. There is no central switchboard that must approve or carry every conversation.

In plain terms: packet switching means the network has no irreplaceable middle. Messages are broken into pieces that each find their own way, so the system keeps working even when parts of it fail. This "no indispensable center" property is the technical seed of the decentralization ideal that Web3 will later claim as its inheritance.

The mathematical theory underpinning packet switching was developed in part by Leonard Kleinrock, who ran the Network Measurement Center at UCLA under a contract awarded in 1968 (Britannica, n.d.). It was from Kleinrock's lab that ARPANET carried its first message, on the night of October 29, 1969, around 10:30 p.m. Pacific time, to the Stanford Research Institute. The operators intended to transmit the word "login," but the receiving system crashed after the first two letters — so the first message ever sent over the ancestor of the Internet was, accidentally, the truncated fragment "lo" (History of Information, n.d.; Britannica, n.d.). The anecdote is charming, but it also captures something true about the early network's spirit: it was an unfinished, failure-prone experiment among a small community of researchers, built to be tinkered with rather than sold.

TCP/IP: the common language that made the Internet

ARPANET was a single network. The Internet, properly speaking, is something more ambitious: an inter-network, a way of joining many separately built and separately run networks into one seamless whole, despite their differences. Achieving that required not just packet switching but a shared set of rules — a protocol — that any network could adopt without surrendering its independence.

A protocol, in computing, is simply an agreed set of rules for how machines exchange information — the digital equivalent of the conventions that let strangers conduct a phone call: who speaks first, how to signal that you did not understand, how to say goodbye.

The protocol that solved the inter-networking problem was described by Vinton Cerf and Robert Kahn in a landmark paper, "A Protocol for Packet Network Intercommunication," published in IEEE Transactions on Communications in May 1974 (Cerf & Kahn, 1974). Their paper introduced what became the Transmission Control Protocol (TCP), designed to let heterogeneous packet networks — networks built by different people, with different hardware and different internal rules — interconnect and exchange data reliably. The design was later split into two cooperating protocols, TCP and the Internet Protocol (IP), known together as TCP/IP.

The cleanest way to grasp the division of labour is again by analogy to the postal system. IP is the addressing-and-delivery layer: it gives every connected machine a numerical address (an IP address, the digital equivalent of a street address) and undertakes to carry individual packets toward that address — but, like a busy postal service handling unregistered postcards, it makes no promise that any given packet will arrive, or arrive in order. TCP is the diligent correspondent built on top of that unreliable delivery: it numbers the postcards, watches for ones that go missing, asks for them to be resent, and reassembles everything into the correct order at the far end. Between them, the two protocols turn a chaotic swarm of independently routed packets into the reliable, ordered streams of data that applications can depend on.

In plain terms: IP is the part that says where data should go and shoves it in that direction without guarantees; TCP is the part that makes sure all of it actually gets there, intact and in order. Together they are the lingua franca of the Internet — the one set of rules that every connected network agrees to speak, no matter how different it is on the inside.

Two features of this design are essential to the thematic argument of this book.

The first is the end-to-end principle. TCP/IP deliberately keeps the network itself simple and "dumb": the routers in the middle do little more than forward packets toward their destination. All the intelligence — error correction, ordering, the logic of particular applications — lives at the ends, in the sending and receiving machines. The network does not need to know or care whether the packets it carries belong to an email, a web page, a phone call, or some application not yet invented. This is what makes the Internet a general-purpose platform rather than a single-purpose service. Anyone can build a new kind of application at the edge without asking the network's operators to change anything in the middle.

The second is that TCP/IP is an open, non-proprietary standard. It was published, not patented for exclusive license; any network could adopt it, and adopting it required no one's commercial permission. When DARPA formally switched ARPANET over to TCP/IP on January 1, 1983, it set the protocol that remains the operational core of the global Internet to this day (ETHW, n.d.). The Internet "won" not because a company marketed it best but because an open standard let an unbounded number of networks bolt themselves together voluntarily.

The decentralized, permissionless design philosophy — and why it matters

It is worth pausing to make explicit a property that the preceding sections have been circling: the early Internet was, by design, decentralized and permissionless. These two words do heavy lifting in the Web3 literature, so they deserve careful definition here, at their source.

A system is decentralized when no single party controls it and no single component is indispensable to its operation. A system is permissionless when anyone can connect to it, use it, or build on it without first obtaining approval from a controlling authority.

The early Internet was decentralized in the sense established above: packet switching gave it no irreplaceable center, and TCP/IP let independently operated networks interconnect as peers rather than as branches of a head office. It was permissionless in a related but distinct sense: because the network was a "dumb pipe" that merely forwarded packets, and because the protocol was open, a researcher or, later, an entrepreneur could attach a new machine or launch a new application without seeking anyone's blessing. No central gatekeeper decided which applications were allowed to exist. This is precisely how the Web itself could later be created — invented by one person, released to the world, and adopted globally — without permission from whoever "ran" the Internet, because in the relevant sense no one ran it.

A frequently cited articulation of this design ethos comes from the engineers who built and chronicled the network. Barry Leiner, Vinton Cerf, Robert Kahn, Leonard Kleinrock and several co-authors, in their widely circulated history "A Brief History of the Internet," describe an architecture deliberately built around open architecture networking, in which constituent networks retain their autonomy and interconnect through a common protocol rather than being subordinated to a central design (Leiner et al., 2009). The Internet, on this account, is constitutively a federation of equals, not a hierarchy.

Two honest qualifications must be entered, because intellectual integrity requires resisting the romanticized version of this history that circulates in promotional Web3 material.

First, "decentralized" was never total. The early Internet had, and still has, important points of coordination and control: the assignment of address blocks and domain names, for example, has always required some central administration (today handled by bodies such as ICANN and the regional registries). The protocols themselves are stewarded by standards organizations. Decentralization in the network's design was an engineering principle about routing and interconnection, not an absolute political condition of no authority anywhere. Conflating the two is a common rhetorical move that this book will scrutinize when it examines Web3's own claims (see Chapter 26, Web3 as Ideology).

Second, the design philosophy was shaped as much by the practical needs of military and academic resilience and by the engineering culture of its builders as by any explicit libertarian politics. The decentralization of the early Internet was, in large part, an engineering solution to the problems of efficiency and fault tolerance. Reading it backward as a deliberate political project of individual sovereignty is anachronistic. What is true, and important, is that the resulting architecture turned out to have political affordances — it made certain kinds of openness, permissionless innovation, and resistance to central control natural — and that later movements, cypherpunk and then Web3, seized on those affordances as an ideal worth defending and extending (see Chapter 7, The Etymology and Intellectual History of "Web3").

In plain terms: the Internet was built so that no one needs anyone's permission to join in or to invent something new on top of it, and so that no single failure or single owner can switch it off. That openness was originally an engineering decision, not a manifesto. But it produced a world in which a new global system — the Web — could be created by one person and adopted by everyone without a gatekeeper's approval. That fact is the historical anchor of the entire decentralization ideal that Web3 later claims as its birthright.

Tim Berners-Lee and the invention of the Web at CERN

By the late 1980s the Internet existed and worked, but using it to find and share information was cumbersome. Different systems stored documents in incompatible ways; retrieving a file meant knowing exactly where it lived and which program to use. Tim Berners-Lee, a British software engineer working at CERN, the European particle-physics laboratory near Geneva, set out to solve this problem of information management for a large, constantly changing community of scientists.

In March 1989 — the document was submitted on March 12 — Berners-Lee circulated a proposal titled "Information Management: A Proposal" (Berners-Lee, 1989). It described a system of linked documents, accessible across the laboratory's incompatible computers, in which any document could point to any other through what we now call hyperlinks. His manager, Mike Sendall, did not immediately fund it but allowed the work to proceed, annotating the cover of the proposal with a phrase that has become famous in the history of technology: "Vague but exciting" (CERN, n.d.; Time, 2014).

A hyperlink is a connection embedded in one document that, when activated, takes the reader directly to another document or resource — the digital equivalent of a footnote that you can step into, instantly arriving at the source it points to rather than merely being told where to find it.

Berners-Lee considered several names for his creation, including "Information Mesh" and "The Information Mine," before settling on World Wide Web (Wikipedia: History of the World Wide Web, n.d.). The system became available as a service on the open Internet on August 6, 1991 (CERN, n.d.). Crucially for everything that follows, CERN placed the Web's underlying software in the public domain in 1993, renouncing any royalty or licensing claim (CERN, n.d.). The Web, like the Internet beneath it, was given away — a decision without which it could not have spread as it did.

The three inventions that constitute the Web

The Web rests on three foundational technologies, all devised by Berners-Lee, that together turn the Internet's raw data-carrying capacity into a navigable space of documents (World Wide Web Foundation, n.d.). Each can be defined plainly.

HTML (HyperText Markup Language) is the language in which web pages are written — a set of simple tags that mark up plain text to indicate what is a heading, what is a paragraph, what is a link, and what is an image. It is, in effect, the standardized grammar of a page, so that any browser anywhere can read and display it.

A URL (Uniform Resource Locator), the most common form of the more general URI (Uniform Resource Identifier), is the address of a resource on the Web — the string beginning with http:// that names exactly one page or file and tells a browser where to find it. It is the Web's equivalent of a postal address, but for documents: globally unique, so that the same address fetches the same resource from anywhere in the world.

HTTP (HyperText Transfer Protocol) is the set of rules a browser and a web server use to talk to each other — the request-and-response conventions by which a browser says "please send me the page at this address" and a server replies with the page. If HTML is the grammar of the page and the URL is its address, HTTP is the conversation that delivers it.

These three fit together in a way the reader uses, unknowingly, dozens of times a day. You click a hyperlink; it contains a URL identifying the desired resource; your browser uses HTTP to request that resource from the server named in the URL; the server returns a document written in HTML; your browser renders it on screen, complete with new hyperlinks to follow. Notice that all of this happens on top of TCP/IP: HTTP's requests and responses are carried as packets across the Internet by the very protocols described earlier in this chapter. The Web is, quite literally, an application layered on the Internet — the clearest possible illustration of the distinction with which this chapter opened.

In plain terms: HTML is what a page is written in, a URL is a page's address, and HTTP is the polite exchange by which your browser asks for a page and gets it back. The Web is just these three conventions agreed upon worldwide — and they ride on the Internet's pipes rather than replacing them.

The original vision: an open, decentralized Web

Berners-Lee's Web inherited and extended the design ethos of the Internet beneath it. It was, by intention, open in several reinforcing senses. Its specifications were public and unowned; CERN's release of the software into the public domain meant no company controlled the standard. It was permissionless in the same way the Internet was: anyone could run a web server and publish a page, and anyone could write a browser to read pages, without applying to a central authority. And it was decentralized in its addressing: because a URL could point to any server anywhere, there was no central index, no master directory, and no single host through which all content had to pass. The Web was conceived as a single, universal information space in which any resource could link to any other across institutional and national boundaries.

This original vision is important to state precisely, because it is the reference point against which both the Web's subsequent evolution and Web3's restorative claims are measured. In the Web's first era, publishing was genuinely distributed: an individual or institution that ran a server was a peer on the network, owning its own content and its own address, linking outward and being linked to without intermediation. The architecture did not require, and did not initially have, the dominant platforms — the search engines, social networks, and app stores — that would later mediate most of what people do online. Those came later, and the story of how the open Web acquired powerful centralizing intermediaries is the subject of the next several chapters (see Chapter 5, Web 2.0: The Read–Write Era and the Platform Economy, and Chapter 6, The Discontents of Web 2.0).

It is worth noting, as a matter of even-handedness, that Berners-Lee himself has remained committed to this decentralization ideal and has pursued it through means quite distinct from blockchain. His later Solid project aims to return control of personal data to individuals through user-controlled personal data stores, and he has pointedly declined to describe that effort as "Web3" (Fast Company, 2022). That the Web's own inventor pursues re-decentralization on a non-blockchain path is itself evidence that "decentralizing the web" is a broader and older aspiration than the specific blockchain-based program that calls itself Web3 — a distinction this book treats as load-bearing and returns to in Chapter 27, Disentangling the Terms.

How "decentralized origins" became a recurring ideal

We can now state the thematic payoff that this chapter exists to set up. The history above furnishes a powerful and recurrent narrative that Web3's proponents invoke almost ritually: the network was born free. The Internet was designed without an indispensable center and without gatekeepers; the Web was given away rather than owned; for a time, to be on the web was to be a peer among peers. On this telling, the open, decentralized, permissionless character of the early Internet and Web is not a historical accident but the network's true nature — a nature subsequently betrayed by the rise of dominant platforms, and one that a properly designed successor ought to restore.

This narrative is rhetorically potent precisely because its factual core is real. The Internet genuinely was architected as a decentralized federation; the Web genuinely was launched as an open, ownerless standard. Web3 does not have to fabricate this history; it has only to inherit it. When advocates describe their project as "re-decentralizing the web" or "returning the web to its users," they are positioning blockchain technology as the heir to a design philosophy that demonstrably existed at the origin (see Chapter 25, Toward a Definition, and Chapter 26, Web3 as Ideology).

But this book's commitment to intellectual honesty requires holding two qualifications alongside the narrative, both of which later chapters develop. First, as noted above, the early network's decentralization was always partial and was substantially an engineering achievement rather than a fully realized political ideal; reading it as a lost utopia overstates what was actually there. Second, and more pointedly, the claim of inheritance is not the same as the fact of inheritance. Whether blockchain-based systems actually deliver the decentralization they invoke — or whether they reproduce, under new names, the very centralization they decry — is an empirical question this book takes up at length (see Chapter 29, Critiques and Skepticism: Hype, Scams, and Recentralization in Practice). The genealogy is genuine; the success of the restoration is contested.

What the present chapter establishes, then, is the baseline. The Internet is a decentralized, permissionless network of networks, built on packet switching and the open TCP/IP standard, on top of which the Web — Berners-Lee's open, public-domain system of HTML, URLs, and HTTP — runs as one application among many. Both were born from, and made possible by, a design philosophy that treated central control as a thing to be engineered around rather than relied upon. That philosophy is the historical wellspring from which the Web3 ideal of decentralization is drawn. Everything that follows in this book — the rise of centralized platforms that seemed to contradict that philosophy, and the blockchain-based movement that claims to vindicate it — is a commentary on this origin.

Synthesis

This chapter has drawn the distinction on which much of the book depends: the Internet is the underlying network of networks, while the World Wide Web is one application running on it. The Internet's defining technologies — packet switching, pioneered in ARPANET from 1969, and the open TCP/IP protocol described by Cerf and Kahn in 1974 and made operational in 1983 — gave it a decentralized, permissionless architecture with no indispensable center and no gatekeeper to entry. Tim Berners-Lee's invention of the Web at CERN (proposed 1989, public in 1991, released to the public domain in 1993) extended that ethos with three open standards — HTML, URLs, and HTTP — and a vision of a universal, decentralized information space.

The thematic significance is this: the "decentralized origins" of the Internet and the Web are not a myth but a documented design philosophy, and they have become the recurring ideal that Web3 invokes when it presents itself as a restoration of the network's true, ownerless nature. The book's central thesis — that Web3 is best understood as a contested socio-technical project bundling blockchain technologies with an ideology of decentralization and digital ownership — begins to take shape here, at the point where a real architectural heritage of decentralization is established as the standard against which later centralization will be judged and the banner under which Web3 will later march. The next chapter turns to what that open Web actually looked like in practice during its first decade — Web 1.0, the read era — and to how its limitations set the stage for the platform economy that would follow.

References cited in this chapter

Berners-Lee, T. (1989). Information Management: A Proposal. CERN. https://info.cern.ch/Proposal.html

Britannica. (n.d.). ARPANET. Encyclopædia Britannica. https://www.britannica.com/topic/ARPANET

Britannica. (n.d.). Leonard Kleinrock. Encyclopædia Britannica. https://www.britannica.com/biography/Leonard-Kleinrock

CERN. (n.d.). A short history of the Web. https://home.cern/science/computing/birth-web/short-history-web

Cerf, V. G., & Kahn, R. E. (1974, May). A Protocol for Packet Network Intercommunication. IEEE Transactions on Communications, 22(5), 637–648. https://www.cs.princeton.edu/courses/archive/fall06/cos561/papers/cerf74.pdf

Engineering and Technology History Wiki (ETHW). (n.d.). Milestones: Transmission Control Protocol (TCP) Enables the Internet, 1974. https://ethw.org/Milestones:Transmission_Control_Protocol_(TCP)_Enables_the_Internet,_1974

Fast Company. (2022). Tim Berners-Lee is building the web's 'third layer.' Don't call it Web3. https://www.fastcompany.com/90807852/tim-berners-lee-inrupt-solid-pods

History of Information. (n.d.). Charley Kline Sends the First Message Over the ARPANET from Leonard Kleinrock's Computer. https://www.historyofinformation.com/detail.php?entryid=1108

Leiner, B. M., Cerf, V. G., Clark, D. D., Kahn, R. E., Kleinrock, L., Lynch, D. C., Postel, J., Roberts, L. G., & Wolff, S. (2009). A Brief History of the Internet. ACM SIGCOMM Computer Communication Review, 39(5), 22–31. https://doi.org/10.1145/1629607.1629613

Time. (2014). Web at 25: Tim Berners-Lee's Amazing Proposal Document. https://time.com/21039/tim-berners-lee-web-proposal-at-25/

Wikipedia. (n.d.). History of the World Wide Web. https://en.wikipedia.org/wiki/History_of_the_World_Wide_Web

World Wide Web Foundation. (n.d.). History of the Web. https://webfoundation.org/about/vision/history-of-the-web/