The metaphor of a ‘network of networks’ commonly used to explain how the Big Internet works only partially describes it. The Big Internet is, in fact, a technical infrastructure, a social organisation and a set of shared standards all at once. And it is much less free and rebellious than people think by Andrea Monti – Initially published in Italian by MIT Technology Review Italy
The popular narrative about the Big Internet tells of a network without borders, free, uncontrollable by governments, where people can do — and be — whatever they want. Born as a dream of freedom, the story continues, the Big Internet has been under attack for some time by executives from all over the world and is being appropriated by Big Tech. The result of this process is the compression of the rights of those who use it, their transformation into “users” and an abrupt transition from dream to nightmare.
The reality is — and always has been — different, starting with the way this infrastructure is named. Internet (with a capital letter) is not a proper noun but a common noun because it identifies a category of “objects”. Therefore, just as we do not say “I watch Television” or “I use a Microwave”, we should not say “I connect to the Internet”. “Internet” is not a “media”, so it is equally wrong to think of content available through the network in terms of “I found it on the Internet”.
“Internet” vs Big Internet
As will become clearer in a few lines, “Internet” — or rather, the internet — is more of a label to identify those networks that use a specific technological and organisational architecture. In this sense, therefore, there are as many internets as one wants to create, from those that operate in small local contexts to those that cover entire cities, countries or significant parts of the world. The latter case is called, this time with a proper name, “Big Internet” and is the subject of this article.
This introduction, halfway between sociology and semantics, is necessary to understand how the key components of Big Internet interact and what happens when a device connects to one of the many nodes in the network.
Protocols and levels of communication
The DNA of an internet network, what makes it different from others, is the protocol, i.e. a set of formal rules that define how connected devices “talk” to each other: message formats, connection sequences, authentication, error control systems. One way to understand what a protocol is is to compare it to Esperanto — a common language that allows individual terminals to communicate even though each one basically uses its own language.
In the case of an internet network, there are many protocols that are grouped under the acronym TCP/IP (Transfer Control Protocol/Internet Protocol).
Protocols operate at various levels (layers) to manage different functions.
At the Network Layer, the Internet Protocol (IP) defines the addresses of the connected endpoints, how to fragment data into packets, how to route them from source to destination, and much more. The IP protocol is important because it assigns a unique “licence plate” to a terminal to make it reachable.
At the Transport Layer, protocols such as TCP (Transmission Control Protocol), which guarantees order, integrity and congestion control, and UDP (User Datagram Protocol), which is lighter but does not guarantee the successful exchange of data, operate.
At an even “higher” level — i.e. “close” to the user — the application layer operates other application-level protocols/services: http(s) manages the display of web content, SMTP manages the forwarding of e-mail, TLS manages the security of communications on open channels, FTP manages file exchange, and so on.
To do their job, protocols must be incorporated into the operating systems and software that run the network and those used by those who connect to it.
Packets and routing
Data managed through protocols are dynamically routed according to the most efficient — or, as we shall see, politically appropriate — path to reach their destination.
Routing works like the transporter in Star Trek: the file is “broken up” and the sending and destination parameters are added to each piece to ensure that all the individual components arrive safely to be reassembled into the original file.
DNS: the domain name system
To connect to a network resource, as mentioned above, all you need to do is know its IP number and enter it into the interface of the software you are using. However, this is highly inefficient because the number of resources to be used discourages the use of memory or more or less automated directories.
The domain name system was designed to solve this problem. Domain names are alphanumeric strings associated with Top Level Domains (TLDs), which can be geographical (.it for Italy) or thematic (.com, .net to denote economic activities or network structures).
So, when a device wants to contact a network resource (e.g. technologyreview.it), it sends a request to a DNS server called a “resolver”, which acts as a kind of map. It converts the domain into the corresponding IP number and returns the address to the requesting terminal, which can then proceed with the connection.
The DNS is a hierarchical structure, whose vertices —Root Servers— are mostly located in the USA and therefore subject to US jurisdiction and regulations. Therefore, whoever technically controls the Root Servers controls everything that depends on them anywhere else in the world.
Telecommunications operators and internet providers
This entire infrastructure does not exist in a vacuum but is maintained by entities that have access to cables and/or frequencies to enable network connections and data transport (telecommunications operators) and by entities that provide services through data centres, ranging from email to hosting — storage of content on behalf of customers — and so on.
In reality, this distinction is very theoretical because, over the years, the technological and commercial model has become verticalised and the difference between internet providers and telecommunications operators is increasingly blurred.
A side effect of this vertical integration is that as these “hybrid” entities and their user bases grow, the choices in terms of protocols, technologies and tools used to deliver services also become influential for those who use different operators. This transforms operators into entities capable of directly influencing both the functioning of the Big Internet and that of the bodies that traditionally constitute Internet Governance.