How Internet Infrastructure Works

Every device that is connected to the internet is part of a network, even the one in your home. F or example, your computer may use a cable or fiber modem to connect to an internet service provider (ISP). At work, your device may be part of a local area network (LAN), but your internet connection is provided by your employer's ISP. Once you connect your computer it becomes part of your employer's network. The ISP may then connect to a larger network. The internet is simply a network of networks.

Large communications companies have their own dedicated backbones, always-on connections to the internet that have enough bandwidth to allow many people to use the connection at the same time. In each region, a company has a local office that connects local homes and businesses to its main network. The amazing thing here is that there is no centralized network. Traffic travels from point to point, and if one computer drops out of the network, the packets that make up a digital file are routed to another computer. Files arrive as expected, and you'd never notice the change in the traffic pattern.

Here's an example. Imagine that Company A is a small firm that has its office network set up with a server and a networked printer. Imagine that Company B is a corporate ISP. Company B builds or leases office space in major cities to store its servers and routing equipment. Company B is so large that it runs its own fiber optic lines between its buildings so that they are all interconnected.

In this arrangement, all of Company A's customers can talk to each other, and all of Company B's customers can talk to each other, but the two companies' networks aren't linked. Both companies can communicate internally, but neither can communicate with the other. Therefore, Company A and Company B both agree to connect to internet access points, or IXPs, in various cities. The two firms' networks can now connect to each other and other organizations using the internet.

This example shows how two companies' networks talk to each other, but these two businesses are just a close-up example showing how their two networks join the vast internet. To get a bird's-eye view of what these interconnected networks look like, take a look at Barrett Lyon's Opte Project, which endeavors to create an evolving map of internet pipelines.

All of these networks rely on IXPs, backbones and routers to talk to each other. What is incredible about this process is that a message can leave one computer and travel halfway across the world through several different networks and arrive at another computer in a fraction of a second!

The routers determine where to send information from one computer to another. Routers are specialized devices that send your messages and those of every other internet user speeding to their destinations along thousands of pathways. A router has two separate but related jobs:

In performing these two jobs, a router is extremely useful in dealing with two separate computer networks. It joins the two networks, passing information from one to the other. It also protects the networks from one another, preventing the traffic on one from unnecessarily spilling over to the other. Regardless of how many networks are attached, the basic operation and function of the router remains the same. Since the internet is one huge network made up of countless smaller networks, its use of routers is a necessity.

The National Science Foundation (NSF) created the first high-speed backbone in 1986. Called NSFNET, it was a T1 line that connected 170 smaller networks together and operated at 1.5 Mbps (million bits per second). IBM, MCI and Merit worked with NSF to create the backbone and developed a T3 (45 Mbps) backbone the following year.

Backbones are internet connections that allow vastly more traffic than the connection from your home to the central office around the corner. In the early days of the internet, only the largest telecommunications companies had the ability to handle that sort of bandwidth.

Today more companies operate their own high-capacity backbones, and all of them interconnect at various IXPs around the world. In this way, everyone on the internet, no matter where they are and what provider they use, can talk to everyone else on the planet. The entire internet is a gigantic, sprawling agreement between people to intercommunicate freely.

Every machine on the internet has a unique identifying number, called an IP address. The IP stands for Internet Protocol, which is one of two protocols that computers use to communicate over the internet. The other is Transmission Control Protocol, and the two are often referred to as one in the phrase TCP/IP. A protocol is the predefined way that someone who wants to use a service connects with that service. The "someone" could be a person, but more often it is a computer program like a web browser.

A typical IP version 4 (IPv4) address looks like this: 216.27.61.137.

To make it easier for us humans to remember, IP addresses are normally expressed in decimal format as a dotted decimal number like the one above. But computers communicate in binary form. Look at the same IPv4 address in binary: 11011000.00011011.00111101.10001001.

Each sequence of numbers in an IPv4 address is called an octet, because each has eight positions when viewed in binary form. If you add all the positions together, you get 32, because IPv4 addresses are considered 32-bit numbers. Since each of the eight positions can have two different states (1 or 0), the total number of possible combinations per octet is 2⁸ or 256. So, each octet can contain any value between zero and 255. Combine the four octets and you get 2³² or a possible 4,294,967,296 unique values!

Out of the almost 4.3 billion possible combinations in IPv4 addresses, certain values are restricted from use as typical IP addresses. For example, the IP address 0.0.0.0 is reserved for machines on the local network and the address 255.255.255.255 is used for broadcasts.

Although 4.3 billion sounds like a lot of addresses, the internet has grown so fast that a newer 128-bit address system was needed to replace IPv4. The experts at the Internet Engineering Task Force (IETF) began working on a new system in late 1998. IP version 6 (IPv6), which officially launched on June 6, 2012, has room for 340 trillion³ addresses, so we should have plenty of room for all our devices. (For now.) Just for the record IPv5 was never formally adopted as a standard.

As you might expect, IPv6 addresses look a little different from IPv4, which was created in the 1970s. Each segment in an IPv6 address uses four numbers and is separated by a colon.

An example looks like this: ba5a:9a72:4aa5:522e:b893:78dd:a6c4:f033.

Because IPv6 uses hexadecimal notation, there are 16 individual digits that need to be represented. So besides the numbers zero through nine, the letters A-F have been drafted to stand in for the double-digit numbers.

Sticking with IPv4 for the moment, octets serve a purpose other than simply separating the numbers. They are used to create classes of IP addresses that can be assigned to a particular business, government or other entity based on size and need. The octets are split into two sections: network and host. The first octet is used to identify the network that a computer belongs to. Host (sometimes referred to as node) identifies the actual computer on the network. The last octet shows the host segment. There are five IP classes plus certain special addresses. You can learn more about IP classes at What Is an IP Address?

When the internet was in its infancy, it consisted of a small number of computers hooked together with modems and telephone lines. You could only make connections by providing the IP address of the computer you wanted to establish a link with. For example, a typical IP address might be 216.27.22.162. This was fine when there were only a few hosts out there, but it became unwieldy as more and more systems came online.

The first solution to the problem was a simple text file called a host table maintained by the Network Information Center (NIC) that mapped names to IP addresses. Soon this text file became so large it was too cumbersome to manage. In November 1983, Paul Mockapetris submitted two requests for comments to the International Network Working Group. RFC 882 outlines concepts of domain name system (DNS), which maps text names to IP addresses automatically. RFC 883 proposes ways of implementing the system. Thanks to his and many others' efforts, this way you only need to remember www.howstuffworks.com, for example, instead of the series of numbers and punctuation that is HowStuffWorks.com's IP address.

When you use the web or send an email message, you use a domain name to do it. For example, the Uniform Resource Locator (URL) "https://www.howstuffworks.com" contains the domain name howstuffworks.com. So does this email address: example@howstuffworks.com. Every time you use a domain name, the internet's DNS servers translate the human-readable domain name into the machine-readable IP address. Check out How Domain Name Servers Work for more in-depth information on DNS.

Top-level domain names, also called first-level domain names, include .COM, .ORG, .NET, .EDU and .GOV. Within every top-level domain there is a huge list of second-level domains. For example, in the .COM first-level domain there are:

Every name in the .COM top-level domain must be unique. The left-most section, like "www," is the host name. It specifies the name of a directory on a specific machine (with a specific IP address) in a domain. A given domain can, potentially, contain millions of host names as long as they are all unique within that domain.

DNS servers accept requests from programs and other name servers to convert domain names into IP addresses. When a request comes in, the DNS server can do one of four things with it:

Let's say that you type the URL www.howstuffworks.com into your browser. The browser contacts a DNS server to get the IP address. A DNS server starts its search for an IP address by contacting one of the DNS root servers. The root servers know the IP addresses for all of the DNS servers that handle the top-level domains (.COM, .NET, .ORG, etc.). Your DNS server asks the root for www.howstuffworks.com, and the root would say, "I don't know the IP address for www.howstuffworks.com, but here's the IP address for the .COM DNS server."

Your name server then sends a query to the .COM DNS server asking it if it knows the IP address for www.howstuffworks.com. The DNS server for the .COM domain knows the IP addresses for the name servers handling the www.howstuffworks.com domain, so it returns those.

Your name server then contacts the DNS server for www.howstuffworks.com and asks if it knows the IP address for www.howstuffworks.com. It actually does, so it returns the IP address to your DNS server, which returns it to the browser, which then contacts the server for www.howstuffworks.com to get a web page.

One of the keys to making this work is redundancy. There are multiple DNS servers at every level, so that if one fails, there are others to handle the requests. The other key is caching. Once a DNS server resolves a request, it caches the IP address it receives. Once it has made a request to a root DNS server for any .COM domain, it knows the IP address for a DNS server handling the .COM domain, so it doesn't have to ask the root DNS servers again for that information. DNS servers can do this for every request, and this caching helps to keep things from bogging down.

Even though they're totally invisible, DNS servers handle billions of requests every day, and they are essential to the internet's smooth functioning. The fact that this distributed database works so well and so invisibly day in and day out is a testimony to the design.

Every machine on the internet is either a server or a client. The machines that provide services to other machines are servers. And the machines that are used to connect to those services are clients. There are web servers, email servers, FTP servers and so on serving the needs of internet users all over the world.

When you connect to www.howstuffworks.com to read a page, you are a user sitting at a client's machine. You are accessing the HowStuffWorks web server. The server machine finds the page you requested and sends it to you. Clients that come to a server machine do so with a specific intent, so clients direct their requests to a specific software server running on the server machine. For example, if you are running a web browser on your machine, it attempts to talk to the web server on the server machine, not the email server.

A server has a static IP address that does not change. A home machine that is dialing up through a modem, on the other hand, typically has an IP address assigned by the ISP every time you log on. That IP address is unique for your session — it probably will be different the next time you dial in. This way, an ISP only needs one IP address for each device, rather than one for each customer.

Any server makes its services available using numbered ports — one for each service that is available on the server. For example, if a server machine is running a web server and a file transfer protocol (FTP) server, the web server would typically be available on port 80, and the FTP server would be available on port 21. Clients connect to a service at a specific IP address and on a specific port number.

Once a client has connected to a service on a particular port, it accesses the service using a specific protocol. Protocols simply describe how the client and server will have their conversation. Every web server on the internet conforms to the hypertext transfer protocol (HTTP). You can learn more about internet servers, ports and protocols by reading How Web Servers Work.

Networks, routers, NAPs, ISPs, DNS and powerful servers all make the internet possible. It is truly amazing when you realize that all this information is sent around the world in a matter of milliseconds! The components are extremely important in modern life — without them, there would be no internet. And without the internet, life would be very different indeed for many of us.

For more information on the structure of the internet and related topics, check out the links that follow.