Article content overview
In the previous article, I learned that when transmitting across devices, you need to use the routing table to know the IP of the next hop of the datagram
There are two columns of important information in the routing table, namely: the destination IP address and the next hop IP address . These two IPs form a mapping relationship. When I learned about datagram forwarding, the routing table is directly used, so there will be several questions
How does the next hop address in the routing table come from?
Is the next hop address unique?
Is the next hop address the best?
With so many routers, how do they coordinate their work?
These problems are the problems that need to be solved in network layer routing, in fact, a good algorithm is needed to solve these problems. Therefore, to understand the routing of the network layer is to understand some of the algorithms inside
A network topology used many times in the previous article. In this topology, there are computers, routers, small networks, medium networks, and large networks.
This picture is actually a thumbnail of the Internet in real life, and it can objectively describe the Internet in the real world. However, when learning the algorithm, this graph cannot intuitively express the problem we need to solve. Therefore, when understanding the algorithm inside, it is converted into a data structure that is easy for the computer to understand. Therefore, this network topology, Usually this network topology is transformed into a graph for description
The picture below is a transformation of a simple network topology, which has six vertices A, B, C, D, E, F and several edges. Each vertex represents a network or router or computer. Each edge represents a network path. For example, there are many paths to go from A to E, and this is also true in real networks. Therefore, the graph can objectively represent the network in real life
If the Internet is abstracted into a graph, then the routing algorithm is actually a graph theory algorithm . However, the routing algorithm is not only a graph theory algorithm, because the Internet is very complicated, for example, the network often jitters or the network fails. This will cause the graph to change in real time. Therefore, the routing algorithm is actually much more complicated than the graph theory algorithm
An ideal routing algorithm has the following characteristics:
- The algorithm is correct and complete (that is, according to the algorithm, the data can reach the destination correctly and completely)
- The algorithm should be as simple as possible in calculation (because there are too many devices in the network, the Internet is very complex and huge. If you run complex algorithms on the huge Internet, it is not advisable)
- The algorithm can adapt to changes in the network (because the state in the network changes in real time, for example, at a certain moment, a router fails, or the fiber or cable is broken)
- The algorithm is stable and fair (stability means that even if a device fails, it will not affect other networks or devices. Fairness means that the algorithm is fair to every device (computer, router), not because of equipment Different, leading to different forwarding paths)
Mentioned before:
- The scale of the Internet is huge
- The Internet environment is very complicated
Therefore, in order to make the routing algorithm simpler, the Internet is also artificially processed. Internet has been divided, which contains a very important concept: an autonomous system (Autonomous System, referred to as AS)
- An autonomous system (AS) is a network equipment group under a management organization (management organization: it can be a country, a city, or a company)
- The internal network of the AS is managed by itself, and the AS provides one or more entrances and exits to the outside (the equipment outside the AS does not directly participate in the internal management, if the external equipment needs to access the AS, the access will be made through one or more entrances provided by the AS)
In the previous article, the network level was divided as follows
After understanding the AS, you can know that, in fact, every backbone ISP can be understood as a backbone AS, and every regional ISP can be understood as a regional AS. After understanding the concept of autonomous system AS, routing algorithms can be divided into two levels:
- The routing protocol running in the autonomous system is called: interior gateway protocol (RIP, OSPE)
- The protocol that runs on the external routing of the autonomous system is called: External Gateway Protocol (BGP)
You can understand the interior gateway protocol and exterior gateway protocol through a diagram
Suppose there is an AS1 (may be a company's network) and an AS2 (may be a school's network). When a company and a school communicate with each other, it may be routed through the external gateway protocol
It is the core competitiveness of a technical person to find the constant in the rapidly changing technology. Unity of knowledge and action, combining theory with practice