Welcome to the first post in our series, “Techsplanations” – our first topic will be the physical structures that make up the internet.
For the following posts in this series, please refer to this glossary as a reference for the key terms (in bold) and concepts involved. If you already have a good grasp on the topics presented, consider using this series as an educational entry point for your friends and relatives. For those who would like to have a better understanding of the technologies that shape our everyday lives, whether to make informed decisions about your interactions with technology or to engage with policy and lawmakers at a more granular level, read on.
What is the internet, anyway?
Although many of us use the internet everyday, not everyone understands what the internet is or how it works. So let’s start with the basics. The internet consists of a global suite of networks that connect to each other and that transmit information using common protocols for how to handle that information. In other words, the internet is a bunch of smaller networks that can talk to each other to form a much bigger network. Although we may cover additional network structures in the future, for now we will focus on networks connected by wires, cables, and fiber (sometimes called “wireline networks”).
Let’s look at the general layout of the internet from where you sit in front of your computer. Whether at home or at work, your computer is most likely connected to the internet either by a cable or wirelessly (Wi-Fi), to a Local Area Network (LAN). This LAN is one of the smaller networks we mentioned above. It may have several other devices (often called “endpoints”) on it besides your computer, such as your phone, your printer, your television, or even your fridge. Each of these devices connects to the rest of the LAN and to other, external, networks through a common node – usually a router.
Moving upstream from the router, you will next encounter the device that connects your home LAN to a larger network: the modem. (In some cases, a single device can handle both the modem and router functions, rather than two devices.) In most cases, your home LAN’s router will connect via modem to your internet service provider’s (ISP’s) network, typically called an access network. The modem acts as a translator of sorts, converting the digital signals from your devices into the right kind of signal to send over whatever connects your home or office to the ISP’s access network, whether that is copper wire, coaxial or fiber optic cable, or radio waves. This link between your modem and the first node in the ISP’s access network is sometimes called the “last mile,” although the distance may be more or less than a mile.
In many ways, the ISP’s access network resembles a much larger version of your home LAN; it consists of many unique endpoints connected to each other and to a larger set of networks through a series of hubs. The networking equipment is larger, as are the distances between nodes, but the structure is similar; your home network is to the access network as a leafy twig is to the tree.
Beyond the overall structure, access networks also perform similar functions to LANs. They carry information to and from endpoints, making sure to deliver packets to the correct address and routing them efficiently to accommodate other traffic on the network.
In addition, the ISP’s access network may be directly connected to other access networks, backbone networks, content delivery networks, and edge providers. The physical junctions between these networks are called interconnection points (this is where the “inter” part of internet comes from) or internet exchange points (IXP). The networks to which your ISP’s access network connect may, in turn, share interconnections with even more adjoining networks and backbone segments.
Wait, what? There are bones?
Where access networks perform routing and distribution of data, backbone networks carry data in bulk between different parts of the internet. As far as anatomical analogies go, backbone networks are probably more like the spinal cord than any skeletal elements, but, you know, pretty close. In terms of function, backbone networks are a bit like interstate highways or overseas shipping routes, providing efficient transfer between geographically distant locations. In fact, backbone networks often provide links between internationally-diverse networks and are crucial to the internet’s ability to function efficiently at a global scale.
Content delivery networks (CDNs) fulfill a third set of functions and are conceptually different from access and backbone networks. Instead of using wires and such to provide physical connections between computers, CDNs consist of geographically distributed servers that cache (store) frequently-requested content. The fact is that many people consuming content online tend to want similar content, so storing a copy of that content closer to the people who want it cuts down on the delivery time and reduces traffic on backbone networks. Whether CDN servers are located alongside access networks, such as at an exchange point, or within the access network, they serve as local repositories for popular content hosted elsewhere on the internet.
So, rather than the original content provider delivering thousands or millions of individual copies of, say, a streamed TV episode to each individual endpoint requesting it, a CDN would store a copy in a server on or near the access network and serve the movie to local endpoints. This shortens the distance and number of network links data must travel, which improves the end users’ Quality of Experience by reducing latency; web pages load more quickly and movies stream more smoothly. Local storage and distribution of content also reduces the traffic load on backbone networks because fewer copies of the same information are needed to distribute content effectively. CDNs work well for serving stored information, but would not be useful for other kinds of traffic like real-time voice and video transmissions; in both of these cases the “content” is being generated in real-time and cannot be stored in a CDN’s cache beforehand like a TV show or movie can be.
So how does everything get to where it’s going accurately?
Just like a postal address or telephone number, a network address tells the network where to send things to. Your ISP assigns a single IP address to your connection point. This is your public network address and all internet traffic sent to you will be directed to this address. For your LAN, however, each device needs its own address so that it gets the correct information stream. The router assigns private network addresses to each connected device and uses network address translation (NAT) to keep track of which device sends and receives what information so that it can route information streams correctly. This information is sent as one or more Internet Protocol (IP) packets, which we will discuss in more detail in another post.