How the internet really works

IP Transit and Tiers of Transit Providers

In this article, we will talk about IP Transit and Tiers of Transit Providers. The Internet is made up of a diverse range of Internet Service Providers (ISPs). These ISPs have two main options for connecting to the Internet: IP transit and peering. IP transit involves purchasing access to the Internet from an upstream provider, while peering involves forming a direct connection with other ISPs.

Many ISPs find that utilizing a combination of both IP transit and peering allows them to achieve the best balance between cost-effectiveness and optimal performance. This is why ISPs often use a combination of both options to connect to the Internet.

Peering

Peering is a type of connection between Internet Service Providers (ISPs) where they exchange routing information and network traffic to provide access to their customers’ networks. In this process, ISPs only exchange customer prefixes, not the prefixes received from their upstream provider.

In a peering relationship, there is no charge for the traffic exchanged between the peers, as they do not need to pay their upstream provider to interconnect their customers’ networks. This is known as “settlement-free peering.” Instead, ISPs pay for the port on the fabric at the public peering point (IXP) or, in the case of private peering, share the cost of the circuit. This reduces the volume of IP transit data, thus reducing the overall cost.

To sum up, Peering is a way for ISPs to exchange network traffic directly with each other, reducing the need for transit services and reducing costs. They only exchange prefixes of their own customers and they only pay for the ports at the IXP or the circuit in case of private peering.

IP Transit

IP transit is a paid service that connects a customer’s network or downstream ISP to the global Internet. Unlike peering, the price of IP transit is determined by bandwidth usage, which is typically measured using the 95th percentile method. The transit provider or upstream provider allows the customer’s traffic to pass through its network so that it can reach all possible destinations on the Internet.

IP transit service operates using the Border Gateway Protocol (BGP), so customers who purchase IP transit must operate their own Autonomous System (AS). As a result, customers receive a full BGP Internet table, which includes:

• Prefixes of other customers of the upstream provider
• Prefixes advertised by ISP peers of the upstream provider
• Prefixes received by the upstream provider from its own upstream providers

Additionally, a transit provider will advertise all of its customers’ prefixes to its peers and to all of its upstream providers.

Tier-1, Tier-2, Tier-3 providers

ISPs are organized into a hierarchy of three levels, respectively: Tier-1, Tier-2, and Tier-3.

Single-homed network

Single-homed network topology is when a customer only has one connection to a single upstream provider. This is the simplest and most cost-effective solution, but it also has its downsides. If the connection or router fails, the customer’s entire internet connection will also fail. This is why many companies prefer to have multiple connections to different providers to ensure reliability.

A famous example of this is the Cogent and Telia dispute which has demonstrated the drawbacks of being single homed.

Multi-homed network

Dual-homed network topology is when a customer has more than one connection to more than one upstream provider. This means that if one of the connections fail, the customer will still have another connection to rely on. However, this type of topology still has a single point of failure, which is the customer’s device.

To improve reliability, we can add another router on the ISP side and connect the customer’s router to the provider’s routers. This way, if one of the ISP routers fails, it won’t affect the customer’s connection. However, the customer’s device will still be a single point of failure. To further improve redundancy on the customer side, we can add another router to the topology.

In simple words, Dual-homed network topology is a way to have more than one connection to an upstream provider to ensure reliability in case one of the connections fails. However, the customer’s device will still be a single point of failure.

Ideal network topology

The ideal network topology would consist of at least two upstream providers and at least two routers, that is one provider per router. This is because having multiple connections to different providers and multiple routers increases reliability and protects against outages. If one provider or router experiences an issue, the other connections and routers can act as a backup.

This means that the customer’s internet connection will remain stable even in the event of an outage.

However, this type of topology is more complex than single-homed or dual-homed topologies and comes with a higher cost. The complexity and costs are worth it, as it provides the best protection against most frequent outages.

Conclusion

In conclusion, understanding ISP interconnection is crucial for organizations to ensure they choose the best solution that meets the business needs. Hopefully you now have a better understanding of IP Transit and Tiers of Transit Providers.

This includes understanding the different options available, such as IP transit and peering, as well as the different topologies, such as single-homed, dual-homed, and multi-homed. It’s important to note that Border Gateway Protocol (BGP) is a very solid protocol but it requires careful and professional planning and execution due to its inherent complexity.

With the correct planning and execution, organizations can make the best use of their resources while ensuring a stable and reliable internet connection.

If you would like to know more, consider visiting the North American Network Operators’ Group (NANOG) website, it contains a wealth of information.