The rise of the commercial Internet and how it grew into the network of networks. Part 2
The Internet is a network of networks. It is a truly decentralized collaborative creation, and in this series of articles, we are explaining how it works.
In Part 1, we took the journey through time and learned how this once-unimaginable idea of interconnected earth came to fruition. The Internet, originating from ARPANET, spread through the world like a virus. Albeit rapid expansion, the Internet for the longest time was government-funded and restricted to non-commercial uses such as scientific research. Only in the 90s, after passing new legislation in the US, precursors of modern-day ISPs started forming, and it kickstarted the commercialization of the Internet.
Then, the real evolution began.
The Internet Backbone
Despite not being available for the general public (with limited exceptions), in the 1990s, the Internet was now in the making for over 20 years. Therefore, physical infrastructure got established to some extent, the network had time for growth, and the foundation for the Internet had been laid. This foundation is what we now call the Internet Backbone.
At that time, networks like NSFNET provided the Backbone for the USA. Europe had NORDUNET, and there were more. After the door opened for the commercialization of these networks, Network Service Providers (NSPs) initially connected to these backbones via the Network Access Points (NAPs). Once commercial use of the Internet gained momentum, NSPs started interconnecting, thus expanding the Backbone, but this time with the initiative and motivations of private businesses.
In 1994 NSFNET commissioned private companies to take direct control of all four of their NAPs, and NSFNET itself became a part of the history not long after. Eventually, these public NAPs became congested. Leading telecommunication companies, who were among the first real ISPs, began building their own Backbone with better, faster access points. Alongside these established enterprises, smaller operators were also providing Internet access and expanding the network. However, to understand how the Internet works, how data travels, and what issues occur from it, one needs to understand several key concepts, and one of them is Peering.
Peering
Peering is an act of interconnection of separate Internet networks to exchange traffic between each other to gain mutual benefit. Usually, neither party pays the other for the traffic, and this is called settlement-free peering. Other benefits include increased capacity, redundancy, and efficiency of the networks.
Once the largest ISPs have established themselves in the market, they worked out peering agreements between each other. At first, they also willingly peered with smaller players, but that didn’t last. During the first decade of commercialization, more prominent providers broke their arrangements with smaller providers. Consequently, they started charging them to transfer data through their networks and thus connect to the Backbone.
Largest ISPs who were joined by settlement-free peering agreements and required no upstream relationships formed the foundation of the Internet Backbone and are called Tier 1 ISPs. The hierarchy started emerging, and this brings us to the other important concept of Internet economics — Transit.
Transit
A transit occurs when one network operator pays another network for Internet access. Tier 1 ISPs that form the very Backbone are at the apex of a hierarchy and never purchase transit agreements from other providers. They do, however, sell their bandwidth to other, smaller players — Tier 2 ISPs. The latter find themselves in the middle of the Internet’s hierarchy. They form transit agreements with Tier 1 ISPs and purchase bandwidth from them. Not only that, but they engage in peering between themselves and also sell bandwidth to the last frontier companies — Tier 3 ISPs, which strictly purchase traffic from higher tier providers.
Internet Exchange Points
As you now can imagine, to make the Internet work and to ensure full access for everyone, all providers must interconnect seamlessly. At this point, Internet Exchange Points (IXPs) come into play. Originating from NAPs, IXPs work similarly. They are the physical points through which Internet Service Providers and Content Delivery Networks (CDNs) exchange Internet traffic. They allow reducing latency and facilitate peering agreements.
Content Delivery Networks
CDNs form a crucial layer in the Internet ecosystem. They consist of many physical nodes in which the most demanded data can be cached and accessed every time there is an inquiry. The logic is quite simple. The closer physically a user is to the server, the faster data travels to him. Also, the fewer nodes it takes for it to reach the end-user, the quicker it can get there. Because of this, CDN’s performance mostly depends on its geographical location.
The Data Flow
To grasp the concept of data flow on the Internet — imagine a tree. The roots, stem, and branches of the tree form the Internet Backbone. The water is traffic and it flows from the soil throughout the tree.
Let’s say that an Internet user connected from his home in suburban Los Angeles loads a video from the website which stores it on the server in Germany. If a webmaster is using the CDN services and that particular page is often requested from the US, most likely that a copy of that video is cached somewhere in North America. Just like the water can be stored in either the soil or the stem of the tree.
However, for this example, let’s say that CDN services are not being used.
Most likely, the user from the suburbs will be connected by the Tier 3 ISP. Just like the user pays his provider for Internet access, the provider itself pays the upstream Tier 2 providers who, in turn, pay the Tier 1 providers. Note that higher tier providers not only buy transit but also peer together. In reality, data flows in a significantly more complex manner. ISPs may have separate connections to upstream providers, and they might be customers of multiple providers.
The video, downloaded from Germany, will travel through the Atlantic Ocean and continental America via high-speed Internet backbone controlled by multiple Tier 1 ISPs. Then it will get into pipelines of Tier 2 ISPs, who will transfer data packets towards the end goal. Only in the last mile connection Tier 3 ISPs will come into play, and the video will reach the user.
Remember the tree analogy. Stem, which consolidates all of the tree, is a bundle of Tier 1 ISPs. Branches have links to the stem, and therefore the water can flow from the soil to the branches. In turn, branches are formed from many smaller branches and connected to the stem via one thicker branch. This is an example of the peering of Tier 2 ISPs. Leaves are the last mile connection and the end goal of the water flow. They are Tier 3 ISPs, and they can only receive traffic from central parts of the network.
This is an oversimplification, but that is basically how the Internet works. To ensure the smoothest flow of the information, IXPs, CDNs, various networking protocols, etc. come into play on multiple layers and points of the Internet infrastructure. However, these processes are optimized in such a way that the data travels in the shortest route possible, which is not necessarily the most efficient one. This is where Syntropy comes into play, and this is why we need your nodes. Without you, we would not be able to gather information from these different layers of networks throughout the world.
With so many interconnected networks, there are even more possibilities for data packets to travel, and in the next article, we will take a more in-depth look into these processes and how Syntropy will be able to improve them.
