Enterprise Network Management: Dynamic Routing Protocol (1)

VPN: Enhanced Interior Gateway Routing Protocol

It was an enhanced version of the early IGRP that added a strong internal gateway routing protocol. It has made some extensions to IGRP, And the principle is similar. The difference is not very big. So we will introduce it here, we will not introduce IGRP any more, and we will not use IGRP any more in the future. However, it is a proprietary routing protocol of Cisco, which actually limits the scope of use of the network, because only Cisco products in the same network can use the network, but consider the expansion of a network. What if we want to use products from other manufacturers in the future? Therefore, it is important to consider using it without the need to use it. Next, let's take a moment to look at the knowledge points related to VPN. First, we will start with the knowledge of the network-based VPN Gateway:

VPN features:

Earlier in the day when IOS9.21 was deployed, the version of the ios sdk was 12.4. You can find the related IOS files on the official website:

▲Ios Software

The key is that the network is a hybrid routing protocol, that is, it combines the characteristics of Distance Vector and link status. However, distance vectors and link states have their own drawbacks. For example, distance vectors are prone to generate loops, and the calculation path based on the number of hops is not very accurate. The Link Status Protocol has no loops and uses cost as the metric value as the optimal route. However, because of its complicated algorithm, it must occupy certain resources. Then, in combination with them, the following features are available:

Fast Convergence: Compared with RIP, the convergence speed is much faster than OSPF, but it is definitely faster than RIP, when we introduced RIP, we also saw that RIP has a 30-second broadcast time, 180 seconds of invalid time, and 240 of refresh time. So when the network changes, this convergence speed is very slow. In this way, the link state convergence mode is used for the external announcement of the chain state, so the VPN gateway is much faster.

Classless routing: it uses an algorithm that was discovered by Dijkstra, a Dutch computer scientist in 1959. This algorithm is a bit complex, however, it can ensure that each vro in the network has an entire network link diagram without generating a network loop.

Incremental Route update: RIP sends the entire route table to the other party, while VPN sends the updated route to the other party. This is similar to what we call Incremental backup and differential backup, therefore, the performance is relatively high. Trigger update is used. If no update is sent, it is different from RIP!

Supports load balancing of equivalent and non-equivalent paths: RIP supports four equivalent load balancing entries. For some manufacturers, it supports six identical cost entries in OSPF load balancing, however, they are all equivalent Server Load balancer instances. However, only mongoshake can achieve non-equivalent Server Load balancer.

Use unicast or multicast instead of broadcast:

Supports multiple network layer protocols: supports IP, Novell's IPX, Apple's AppleTalk, and so on, mainly because it has a protocol-related module!

Other features are not described in detail, such as support for VLSM and CIDR and manual summarization of supported routes.

The following three tables are shown in the figure:

If you want to work properly, you must use three tables: Neighbor table, topology table, and route table. This is similar to the link status! Let's take a look at how these three tables work collaboratively ,:

▲Three tables of the network in the source China site

The neighbor table stores the ing between the netxt-hop router and the interface, that is, the next hop and the Local interface are stored. Which of the local interfaces can be used to reach the next hop, however, it must be the information of the directly connected route that runs the OSPF router.

Topology table: All route information learned from the neighbor table, that is, information of all target networks, including successor route entry and feasible successor (FS) Route entry, of course, there is only a successor route in the route table. What is the use of this fs? it means that if the successor route fails and the route fails, the feasible successor route will be placed in the route table, it is actually a Redundant Error mechanism!

Route table: stores the best route information to the other party.

Whether a data packet can be forwarded depends on the route table.

Let's take a look at two concepts: FD and AD: these two concepts are also easy to understand:

FD: Feasible Distance

AD: Advertised Distance

What is the relationship between the two terms? For each vro, the FD for reaching each other is calculated based on the advertised distance AD and cost. Then the smallest FD is the best route to the other party.

▲Fd, AD

Someone asks how the metric value is calculated. here we need to refer to the following parameters:

Bandwidth, latency, reliability, load, and MTU must be used here

The formula is as follows: 256 * min (BW) + sum (DLY). If we do not multiply the value by 256, It is the metric Calculation Method of IGRP. Because IGRP will not be used in the future, we will not introduce it here.

BW is the bandwidth, and DLY is the latency.