Comparison of Default and Autonomous routing
The massive functional advantages of Autonomous Internet are founded upon the routing technology Autonomous Routing, which has dramatic advantages over traditional Default Routing. This page compares both technologies.
Most end users use Default Routing, which means they send (default) all their outgoing Internet traffic to a provider, and that provider takes care of routing the traffic to its destination on the Internet.
Autonomous Routing essentially means you make your own routing decisions, completely independent of any provider or carrier. This is exactly what providers and carriers themselves do: they also use Autonomous Routing instead of default routing.
The advantage of default routing is that it is simple. You leave all technical details and routing decisions to your provider. If anything is wrong or not working with regards to routing, you call your provider and rely on him to resolve the problem.
For organizations with small traffic volumes and where Internet is not mission critical for core business, default routing is a simple and appropriate solution.
Default routing does however have serious disadvantages, which generally become more pressing when Internet is core business for your organization and/or you have a higher traffic volume:
- Your Internet connection is inherently not redundant. As you send all of your traffic (default) to one provider, your are completely dependant: if your link to him is down or if their is any problem in his network or his routing, you cannot automatically let your traffic switch over to a backup.
- All traffic is handled at one (high) price.
It is impossible to distinguish between local (national, peering) and global (international, transit) traffic. The cost-price for local traffic is generally much lower then for global traffic. But because your provider cannot measure, let alone predict, where your traffic is going, he will bill all of your traffic based on the wost-case scenario: that all traffic is (expensive) global traffic.
- Indirect routing: less efficient paths
Your provider has full control over how your traffic is routed, and he might decide to route your traffic over less efficient and longer paths, for example because of lower costs or commitments he has made to upstream carriers. He might even oversubscribe (sell more bandwidth then he buys) or there might be inherent bottlenecks somewhere in his network design that you cannot see and cannot control.
- Locked in: Provider Dependent IP Addresses.
You generally get your IP Addresses from your provider, from a larger block (called Provider Aggregatable or "PA" space) that can only be routed on the Internet by him. Therefore, anytime you change provider, you will need to renumber your network.
With Autonomous Routing, decisions how to route your traffic are made one step earlier, at your own router under your control instead of at your provider's router. This ability enables you to:
Implementing and running an Autonomous Network is definitely not rocket science. But it is not as trivial either: it inherently brings more complexity then default routing.