Advanced Routing

When your network becomes really big, or when you start thinking about becoming part of the Internet, you need to be able to dynamically adjust your routing tools. Sites are often interconnected by more and more links. OSPF and BGP4 have almost become the actual standards of the Internet. All supported by gated and Zebra,linux. Because these contents are not included in this document for the time being, we only give specific learning directions here: ciscohttp://www.aliyun.com/zixun/...

A network bridge is a device that is installed on a network and does not require any subsequent configuration. The network switch is basically a multiple-port network bridge. This means that the bridge is a two-port switch. Linux can support a network bridge with multiple interfaces and become a real exchanger. Bridges are often used to improve networks that http://www.aliyun.com/zixun/aggregation/7317.html "> work poorly but cannot be rebuilt." Because the bridge is a 2-layer device (IP below the layer), ...

This chapter contains a number of "convenience recipes", hoping to help you solve practical problems. Convenience recipes cannot replace the principle of understanding, so you should still focus on the inner truth. 1. Run multiple Web sites with different SLAs. There are a number of ways you can achieve this. Http://www.aliyun.com/zixun/aggregation/14417.html ">apache Some of the modules can support this feature, but we'll show you how Linux handles the problem and can provide ...

1. http://www.aliyun.com/zixun/aggregation/9485.html ">ipv6 tunnel" This is another application of the Linux tunneling capability. This is very popular in the early implementations of IPV6. The following hands-on example is certainly not the only way to implement a IPv6 tunnel. However, it is a common method of building tunnels between Linux and Cisco routers that support IPV6, as has been proven by most people. It's probably right for you to ☺. Jane ...

Linux has 3 kinds of tunnels. They are: IP-in-IP tunnels, GRE tunnels, and non-kernel tunnels (such as PPTP). 1. Some comments about the tunnel tunnel can be used to implement many very unusual and interesting features. But if you have a problem with your configuration, there will be terrible mistakes. Do not point the default route to a tunneling device unless you know exactly what you are doing. Furthermore, the tunneling increases the protocol overhead because it requires an additional IP header. Generally it should be 20 bytes per packet, so if the MTU of a network is 1 ...

The kernel has many parameters that can be adjusted in different environments. Typically, the default defaults can meet 99% of the environmental requirements, we don ' t call the http://www.aliyun.com/zixun/aggregation/3415.html ' >advanced HOWTO for The fun of it! There is a very interesting place:/proc/sys/net, you should see. Here in the beginning did not put all the ...

Hopefully this document will help and inspire you to better understand the linxs2.2/2.4 route. What most users do not know is that the tools you use can actually do quite a lot of work. Route and Ifconfig, for example, actually secretly invoke the very powerful Iproute 2 underlying basic functionality. What Linux can do for you a small list: • Control the bandwidth of a particular computer • Control the bandwidth to a particular computer • Help you to share bandwidth equitably • Protect your network from Dos attacks • Protect I ...

As explained in a section of the classification queue, the filter is categorized with the packet and placed in the corresponding child queue. These filters are called within the queue rules of the classification. Here is the classifier (part) We can use: FW determines how the packet is labeled by the firewall. If you don't want to learn TC's filter syntax, this is a shortcut. For details, see the chapter on queues. U32 based on the various fields in the packet, such as the source IP address, and so on. Route is judged by which route the packet will be routed. RSVP,...

So far we have learned how Iproute works, and have mentioned netfilter many times. Here, you can just take the opportunity to see Rusty's notoriously unreliable guide. NetFilter itself can be found here. NetFilter allows us to filter packets or tamper with the packet header. A special feature is that we can put a digital tag on the packet. You can use the--set-mark mechanism. For example, this command has all the packets sent to 25/TCP (outgoing mail) ...

There are several ways to achieve this function. One of the simplest and most straightforward methods is "TEQL"-True (or "normal") link equalization. Like most things done with queues, http://www.aliyun.com/zixun/aggregation/13996.html "> Load balancing needs to be implemented in both directions." Both ends of the link must be involved to have a complete effect. Imagine the following: A and B are routers, and we certainly assume that they are all Linux machines. If you send from Network 1 to ...

Now, when I figure it out, I'm really surprised. Linux 2.2/2.4 can manage bandwidth as well as the highest-end dedicated bandwidth management systems. Even better than frame Relay and ATM. To avoid conceptual clutter, TC uses the following rules to describe bandwidth: mbps = http://www.aliyun.com/zixun/aggregation/12560.html ">1024 kbps = 1024 * 1024 bps = ...

Multicast-howto is already very old (relatively), and inaccurate and even mistaken for 21136.html "> Reader. Before you start multicast routing, you need to reconfigure your Linux kernel to support the type of multicast routing you want to implement. This step requires you to decide what type of multicast route to use. There are basically four kinds: DVMRP (multicast version of the RIP Unicast Protocol), MOSPF (similarly, just OSPF), PIM ("Pr ...").