Technology

How Does the Internet Work?

By BodaWiki EditorialPublished 4 min read

In short

The internet is a network of networks that moves data in packets.[1][2] Each packet carries source and destination addresses, and routers pass it along toward its destination.[2] TCP checks that the data arrives and resends anything lost, and the Domain Name System links names such as icann.org to numeric IP addresses.[3][6]

A long aisle in a data center lined with tall black server racks, with bundles of blue and yellow network cables running along the racks and small status lights glowing.
AI-generated illustration

Key facts

First ARPANET node
UCLA, September 1969[1]
First message sent
"lo" (UCLA to SRI), October 29, 1969[4]
ARPANET switch to TCP/IP
January 1, 1983[1][5]
IPv4 address size
32 bits[5]
IPv6 address size
128 bits[5]
DNS root servers
13 named authorities, run on hundreds of servers worldwide (IANA)[9]
International data and voice carried by seabed cables
Over 95% (NOAA, citing 2009 UN data)[8]
Coordinating bodies founded
IETF 1986; ICANN 1998[6][7]

How is data sent across the internet?

The internet is not a single network but many independent networks joined together, beginning with the ARPANET and soon including satellite and radio networks.[1] Data crosses it as packets: blocks of data that the Internet Protocol (IP) standard calls datagrams.[1][2] IP moves each datagram from a source computer to a destination computer, both identified by fixed-length addresses, and treats every datagram as an independent item with no fixed connection between the two ends.[2]

This is called packet switching.[1][2] The older approach, circuit switching, linked networks at the circuit level.[1]

The networks are joined by devices first called gateways and later routers.[1] A router reads the addresses in a packet's header and makes a routing decision, picking the next network or gateway on the way to the destination.[2] In the original design, gateways were kept simple by storing no information about the individual flows of packets passing through them.[1] Each packet also carries a "time to live" value, set by the sender and reduced along the route; if it reaches zero before the packet arrives, the packet is destroyed.[2]

Between continents, most traffic does not go through satellites but along the seabed.[8] NOAA, citing a 2009 United Nations Environment Programme report, says more than 95 percent of international data and voice transfers are routed through fiber-optic cables on the ocean floor.[8] The first transoceanic fiber-optic cable, linking the United States, the United Kingdom and France, was installed in 1988.[8]

What is an IP address?

Computers on the internet find one another through numbers called IP addresses, and ICANN explains that each address must be unique so computers know where to find each other.[6] Two versions are in use.[5]

  • IPv4, first deployed on 1 January 1983 and still the most common, uses 32-bit addresses often written as four numbers separated by dots, such as 192.0.2.53.[5]
  • IPv6, whose deployment began in 1999, uses 128-bit addresses written in hexadecimal, such as 2001:0db8:582:ae33::29.[5]

Addresses are handed out in layers.[5] The Internet Assigned Numbers Authority (IANA) gives pools of unused addresses to five Regional Internet Registries, internet service providers get theirs from a local, national or regional registry, and providers assign addresses to users.[5]

Regional registryArea covered
AFRINICAfrica[5]
APNICAsia/Pacific[5]
ARINCanada, the USA and some Caribbean islands[5]
LACNICLatin America and some Caribbean islands[5]
RIPE NCCEurope, the Middle East and Central Asia[5]

How do TCP and IP work together?

TCP/IP began as a single protocol, TCP, that handled all transport and forwarding, but it was later split in two: a simple IP that only addresses and forwards individual packets, and a separate TCP that deals with flow control and recovering lost packets.[1] IP on its own is not reliable: it has no acknowledgments, no retransmissions and no flow control.[2]

TCP adds that reliability, giving applications an in-order stream of data.[3] Each piece of TCP data, called a segment, travels inside an IP datagram, and TCP detects lost packets with sequence numbers and errors with checksums, then fixes both by sending the data again.[3]

How does DNS turn website names into numbers?

Long lists of numbers are hard for people to remember, so the Domain Name System (DNS) links names made of letters to the numeric addresses computers use.[6] Early on, a single table listed every host's name and address; once there were many independently managed networks, Paul Mockapetris of USC/ISI invented the DNS as a scalable, distributed replacement.[1]

In a name such as icann.org, the part to the right of the dot ("org", like "com" or "net") is the top-level domain, and one operator, called a registry, is in charge of all names ending in each one.[6] At the top of the system are the root servers, which all store a copy of the same file listing where each top-level domain's own address book can be found.[6] IANA describes them as hundreds of servers in many countries, set up as 13 named authorities.[9] Because computers keep a top-level domain's address once they learn it, ICANN says the root servers are consulted fairly rarely.[6]

Who runs the internet?

The Internet Engineering Task Force (IETF), founded in 1986, develops the internet's standards and publishes them as RFCs, short for "Requests for Comments", which describe technical foundations such as addressing, routing and transport.[7] Its standards are voluntary, and the IETF says it does not control, or even patrol, the internet.[7]

ICANN, formed in 1998, is a not-for-profit organization that coordinates the internet's unique identifiers worldwide, including domain names and IP addresses.[6] It does not control content on the internet, cannot stop spam and does not deal with internet access.[6] IANA handles the global coordination of IP addresses and of the Autonomous System Numbers used for routing internet traffic.[5]

How did the internet begin?

According to the Internet Society, Leonard Kleinrock at MIT published the first paper on packet switching theory in July 1961, and Lawrence G. Roberts, after joining DARPA, published his plan for a network called the ARPANET in 1967.[1] In September 1969 the company BBN installed the first packet switch, called an Interface Message Processor, at UCLA, and the first host computer was connected.[1] On October 29, 1969, a UCLA team tried to log in to a computer at SRI: the letters "l" and "o" got through before the system crashed, so the first message was "lo".[4] By the end of 1969, four host computers were linked.[1]

In the spring of 1973, Bob Kahn asked Vint Cerf to help design a new protocol, and together they worked out what became TCP/IP.[1] On January 1, 1983, the ARPANET switched from its original protocol, NCP, to TCP/IP in a "flag day" change that required every host to convert at once.[1] By 1990, when the ARPANET was shut down, TCP/IP had pushed aside most other wide-area network protocols worldwide.[1]

Frequently asked questions

Who owns the internet?

No single network makes up the internet: it is many independent networks joined together.[1] The IETF writes voluntary standards and says it does not control, or even patrol, the internet, while ICANN coordinates names and numbers but does not control content or deal with internet access.[7][6]

What was the first message sent over the internet?

On October 29, 1969, a UCLA team supervised by Leonard Kleinrock tried to log in to a computer at SRI over the ARPANET.[4] Only the letters "l" and "o" arrived before the system crashed, so the first message was "lo"; the full login worked about an hour later.[4]

What is the difference between IPv4 and IPv6?

IPv4, first deployed in 1983, uses 32-bit addresses and is still the most commonly used version.[5] IPv6, whose deployment began in 1999, uses 128-bit addresses written in hexadecimal.[5]

What happens if a packet gets lost?

IP itself does not recover lost data: it has no acknowledgments and no retransmissions.[2] TCP detects missing packets using sequence numbers and sends the data again.[3]

How does internet data cross the ocean?

Mostly through fiber-optic cables on the seabed rather than satellites.[8] NOAA, citing a 2009 United Nations report, says more than 95 percent of international data and voice transfers travel through these cables.[8] The first transoceanic fiber-optic cable, linking the United States, the United Kingdom and France, was installed in 1988.[8]

Sources

  1. [1]Brief History of the Internet · Internet Society · accessed October 5, 2026
  2. [2]RFC 791: Internet Protocol · IETF / RFC Editor · accessed October 5, 2026
  3. [3]RFC 9293: Transmission Control Protocol (TCP) · IETF / RFC Editor · accessed October 5, 2026
  4. [4]The Day the Infant Internet Uttered its First Words · UCLA (Leonard Kleinrock) · accessed October 5, 2026
  5. [5]Number Resources · Internet Assigned Numbers Authority (IANA) · accessed October 5, 2026
  6. [6]What Does ICANN Do? · ICANN · accessed October 5, 2026
  7. [7]Introduction to the IETF · Internet Engineering Task Force (IETF) · accessed October 5, 2026
  8. [8]Submarine Cables · National Oceanic and Atmospheric Administration (NOAA) · accessed October 5, 2026
  9. [9]Root Servers · Internet Assigned Numbers Authority (IANA) · accessed October 5, 2026
Cite this article
APA
BodaWiki Editorial. (2026, October 8). How Does the Internet Work? BodaWiki. https://bodawiki.com/wiki/how-does-the-internet-work
MLA
“How Does the Internet Work?” BodaWiki, 8 Oct. 2026, bodawiki.com/wiki/how-does-the-internet-work.
Chicago
BodaWiki Editorial. “How Does the Internet Work?” BodaWiki. October 8, 2026. https://bodawiki.com/wiki/how-does-the-internet-work.
  • internet
  • networking
  • tcp/ip
  • dns
  • ip addresses
  • arpanet