Talk about IPv6-some technical details about typical features

Talk about IPv6-some technical details of typical features 0. A huge address space www.2cto.com is no longer needed. In addition to huge space, fixed-length prefix allocation also makes Address allocation more balanced. 1. ICMPv6 is truly useful: The Address/route configuration automation www.2cto.com IPv6 makes the network device plug-and-play, and everything starts with an automatically generated local link address, without this start, there will be no subsequent automation. Any device that supports IPv6 will generate a local link address for the interface. All local link addresses start with FE80. The last 64-bit link address is generated through the link layer address, A typical EUI-64 ing is to map a 48-bit Ethernet MAC address to an interface ID of a 64-bit link local address. After this link local address is generated, even if there is no router, devices on the same link can also communicate with the IP layer through this address. In IPv6, the link layer and the IP layer are cleverly combined. The link layer provides information, and the IP layer uses it to make communication possible. However, the IPv6 protocol stipulates that the local address of the automatically generated link cannot cross the router. We know that the so-called Internet is connected by a router, and Internet data communication must cross the router, that is to say, there must be an IPv6 address that is not a local link address. Since the local link address is available for the same link, it means that the local link address can communicate with the vro of the same link (each interface of a router represents a link ), IPv6 defines a lot of neighbor discovery packets, so that the router can tell the terminal device interface information to be configured. This is called automatic configuration. from the beginning to the end, no manual intervention is required, ordinary users no longer need to worry about the technical work of configuring addresses. The following is the flowchart of automatic configuration: www.2cto.com

The Automatic route push and automatic Address Configuration of OpenVPN are similar to those of IPv6. 2. Address Resolution for ARP and IPv6 in Linux: The status chart is similar to that for IPv6 ethernet address resolution using multicast instead of broadcasting. This process is very simple. If you know the IPv6 address of the node, the multicast address can be automatically calculated. Then, you can send a Resolution Packet request to the multicast address, which is actually discovered by a neighbor, after the response to the MAC-IP on the cache, The MAC-IP on the cache has a state map, the state map a strong look like the Linux ARP cache state map, they are too like, it is simply:

Comparison of the Linux implementation of ARP cache aging time principle analysis, the figure in the http://www.bkjia.com/ OS /201211/167971.html. However, unlike the ARP status chart maintained by Liunx, the adjacent node cache of IPv6 is no longer maintained by complicated computing and non-protocol principles, but by the packets themselves, for example, the update mechanism of ARP cache is to overwrite an arp packet from a device with the same IP address. IPv6 uses the "Overwrite mark" in the adjacent contact announcement message to determine whether to update the packet. The difference between the two is implementation, and the other is the built-in protocol standards. 3. various IPv6 discovery mechanisms: the foundation of automatic configuration IPv6 can be automatically configured. It is because of the existence of many discovery packets that are typically. adjacent contact request/announcement B. router requests/announcements all these discovery packets are carried by ICMPv6. The router discovery mechanism is optional in IPv4 implementation, but it is necessary in IPv6 implementation. The automatic configuration information of addresses and routes is obtained from these discovery packets. Many automatic configurations also require the corresponding automatic failure mechanism. This introduces an IPv6 address-oriented concept, which is also an address attribute, that is, the address status, an address uses one of the following States: Test, valid (select/discard), invalid, no doubt, the specific status depends on the timeout time and received message.
The router configuration is not automated because it needs to indicate the policy of the person, and the host configuration is automated because it is only used. IPv6 routers cannot use the automatic configuration mechanism. Most of the people who configure the router are professionals. The difficulty of IPv4 configuration lies in the routing configuration, the technical work of IP Address Configuration is handed over to the user rather than the Administrator. You must know that the administrator cannot configure every terminal. Although DHCP can, it is a peripheral protocol, not an IPv4 protocol. IPv6 has completely liberated users. With IPv6, all terminals become plug-and-play, similar to traditional household appliances. Therefore, the backbone of IPv6 is configured, while the leaf is self-developed. It is easy to graft the trunk, but the leaves can only grow out. IPv4 leaves are grafted, which is not successful. This is also in line with our common sense. We usually graft a series of bare branches on existing trees, and then wait for the leaves to grow... 4. when you configure a Linux address or route, you may notice a scope parameter. Although you almost never use it, IPv6, this scope is integrated into the IPv6 address itself. Different prefixes indicate different scopes. When using Windows, when you use multiple default gateways configured with multiple NICs or DHCP to obtain multiple default gateways, you may notice a DeadGWDetectDefault in the registry, it determines which default gateway to use based on the efficiency of TCP connections. However, in IPv6, the periodical router announcement has become a standard, and the protocol itself has done a good job of DeadGWDetect and reduced misjudgment, better. In section 2nd, I explained the Linux arp implementation and IPv6 address resolution. This is also an idea that will eventually become a standard. Maybe many protocol stacks that are not open source are also implemented in this way. Therefore, IPv6 can be said to be a collection of everyone's protocol stacks and has solved many problems. 5. sorry about strong and weak hosts. 6. IPv6 will bring about the fundamental transformation of the application mode. Currently, a large number of applications are in the C/S mode (B/S is also a type of CS), and the data stream features are basically unidirectional, the access layer of the WAN also caters to this situation. ADSL is the embodiment, and asymmetry is caused by IPv4 defects and NAT repair. After the popularity of IPv6, all nodes in the world will be connected and interconnected theoretically, and CS applications will remain in the status quo. A large number of P2P applications will emerge, and the interaction of the entire Internet will be greatly enhanced. In addition, IPv6 will affect the WAN Access Technology. ADSL will not be popular any more. We would rather leave it alone. In addition to applications, routers, and other hardware devices, NAT is no longer a required feature, and the router is actually a router. We are now using home routers such as TP-LINK, D-LINK, or even syslink is not what a router, but an Internet access equipment, do not know why the original for this thing named as a router. A pair of vro ports should be interconnected in two directions. IPv6 makes NAT no longer popular. Many home router manufacturers are estimated to have to make major upgrades to their own software.