We're going to take a look at some important concepts around how different autonomous systems relate to each other. These terms reflect the business and commercial relationships between companies, and we use BGP to make the networks implement these commercial agreements. Firstly, there's transit. Transit is a service where a network operator allows traffic to pass through their network to the rest of the Internet. This is a chargeable service. A transit customer pays the network for this service, and the network may in turn need to buy transit from another larger network to have connectivity to the whole internet. So for example, the end users of an ISP are transit customers. They pay the ISP a fee. Part of this fee covers the cost of their access circuit, but the fee also includes the cost of transit, which means, the cost of carrying their traffic across the ISP zone backbone, and the cost of using the ISP's upstream transit connections to reach the rest of the internet. Larger ISPs can sell transit to smaller ISPs. The smaller ISPs can choose who to buy transit from. They may buy transit from multiple providers for extra capacity or for redundancy. And they choose their transit providers based on different criteria, such as cost, reliability, and which transit providers are available in particular locations. Transit is very much a supplier, consumer relationship. The other main relationship is peering. Peering means an exchange of traffic which is mutually beneficial, and since the benefit is roughly equal, to the both parties, there's no exchange of money or settlement free. Let's have a look at a diagram to illustrate these concepts. At the top half of the diagram, in the middle, there's a large backbone provider 'C', which is selling transit to smaller ISPs, 'A' and 'E' in different regions. And on the bottom half of the diagram the large backbone provider 'D' is selling transit to smaller ISPs 'B' and 'F'. Let's think about things from the point of view of provider 'A'. Provider 'A' sells transit to its own customers, which aren't shown, imagine them hanging off the provider 'A' cloud. In order to provide this service, 'A' in turn buys transit from 'C'. That is, 'C' agrees to carry 'A's traffic to the rest of the Internet in exchange for money. Provider 'A' can reach all of 'B' 'C' 'D' 'E' and 'F' through their link to 'C'. It's the job of 'C' to work out how to do this for them. Similarly 'B' is buying transit from 'D' to reach the rest of the internet. Now let's say 'A' and 'B' are geographically close to each other. If there's a reasonable amount of traffic from provider 'A's customers to provider 'B's network, and vice versa, then the two providers can decide to peer with each other, as a local internet exchange point. This means that the traffic between 'A' and 'B' flows directly saving them both money on their respective transit links. But it's very important that the agreement between 'A' and 'B' is enforced. For example, 'A' should not be able to route traffic through network 'B' to the rest of the internet, they would effectively be stealing 'B's upstream bandwidth. So we use BGP to implement the agreed policy. Which says that traffic from 'A' can reach 'B's network, by the exchange point, but cannot transit through 'B' to reach anything else; and the same from 'B' to 'A'. In the middle, are two large providers. 'C's customers include 'A' and 'E', and 'D's customers include 'B' and 'F'. If these providers are of similar size, and so it's of equal benefit to them, they may peer with each other, exchanging traffic between all their customers networks at no cost. So for example, traffic from 'A' to 'F' may go through this peering link with no settlement between 'C' and 'D'. In practice most backbone providers have to buy some transit to guarantee that they can reach all of the internet. There is a very small number of large networks at the top, known as Tier 1s, who peer with each other and can reach the whole internet with buying transit from anyone else. Right now this is over 700,000 prefixes and continues to grow. The parts of the internet which work this way and don't have any default route, are called the "default free zone", and this is what you could think of as the core of the Internet as opposed to the edge. So that's an overview of the key commercial relationships in the internet, which are implemented using BGP. Transit allows traffic to pass through your network to reach the rest of the Internet, and so is a service that is sold. Peering allows traffic which originates and terminates in two adjacent networks to be exchanged and takes place without exchange of money. Single homed transit customers can use a default route, but inside the default free zone you need to carry all of the internet prefixes in your routing table.
© Produced by Philip Smith and the Network Startup Resource Center, through the University of Oregon.
Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)
This is a human-readable summary of (and not a substitute for) the license. Disclaimer. You are free to: Share — copy and redistribute the material in any medium or format Adapt — remix, transform, and build upon the material The licensor cannot revoke these freedoms as long as you follow the license terms. Under the following terms: Attribution — You must give appropriate credit, provide a link to the license, and indicate if changes were made. You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use. NonCommercial — You may not use the material for commercial purposes. No additional restrictions — You may not apply legal terms or technological measures that legally restrict others from doing anything the license permits.