Address Configuration in IPv6

We know that manual configuration of Host IP addresses is a very complicated task, and managing the static IP addresses assigned to the host is a difficult task, especially when the Host IP address needs to be changed frequently. In IPv4, the Dynamic Host Configuration Protocol, DHCP) Enables automatic settings of Host IP addresses and related configurations. A DHCP server has an IP address pool. The host obtains the IP address from the DHCP server and obtains other related information, such as the default gateway and DNS server, so as to automatically set the Host IP address.

An important goal of IPng is to support "plug-and-play". That is to say, a node can be inserted into the IPv6 network and started in the network without any manual intervention. To this end, IPv6 uses two different mechanisms to support plug-and-play network connections. Example of the first mechanism is the startup Protocol BOOTstrap Protocol, BOOTP), and Dynamic Host Configuration Protocol DHCP ), these two mechanisms allow IP nodes to obtain configuration information from special BOOTP servers or DHCP servers. However, these protocols adopt the so-called "automatic state configuration", that is, the server must maintain the state information of each node and manage the saved information.

The problem with automatic Status Configuration is that you must maintain and manage special automatic configuration servers to manage all "statuses", that is, allowed connections and information about the current connections. The system is acceptable for organizations that have sufficient resources to establish and maintain servers. However, small organizations that do not have these resources do not work in a poor situation.

In addition to automatic Status Configuration, IPv6 also uses an automatic Configuration service called Stateless automatic Stateless Auto Configuration. RFC1971 describes the stateless automatic configuration of IPv6. Stateless automatic configuration requires that the local link support multicast, and the network interface can send and receive multicast.

In the stateless automatic configuration process, the node must take the following steps: first, the node for automatic configuration must determine its local link address; then, it must verify the uniqueness of the local link address on the Link; finally, the node must determine the information to be configured. This information may be the IP address of the node, other configuration information, or both. If an IP address is required, the node must determine whether to use the stateless automatic configuration process or the stateful automatic configuration process.

Specifically, during stateless automatic configuration, the host first attaches its Nic MAC address after the local address prefix 1111111010, the generated local Unicast address IEEE has changed the MAC address of the NIC from 48 bits to 64 bits. If the MAC address of the NIC used by the host is still 48 bits, the IPv6 NIC Driver converts the 48-bit MAC address to a 64-bit MAC address according to the IEEE formula ). Then the host sends a request to the address called neighbor discovrey discovered by a neighbor to verify the uniqueness of the address. If the request does not receive a response, it indicates that the local single point broadcast address set by the host is unique. Otherwise, the host uses a random interface ID to form a new local single-point broadcast address. Then, with this address as the source address, the host broadcasts a configuration request called router request router solicitation to all the routers in the local link, the vro responds to this request with a vro that contains a globally aggregated spof address prefix and other related configuration information. The host uses it to add its own interface ID from the global address prefix obtained by the router, automatically configure the global address, and then it can communicate with other hosts in the Internet.

If no vro serves nodes on the network, that is, the local network is isolated from other networks, the node must look for a configuration server to complete its configuration; otherwise, the node must listen on the router announcement message. These packets are periodically sent to multicast addresses of all hosts to specify configuration information such as network addresses and subnet addresses. The node can wait for the notice of the vro, or send a multicast request to the multicast address of all vrouters to send a notice to the vro. Once the router response is received, the node can use the response information to complete automatic configuration.

With stateless automatic configuration, you can change the IP addresses of all hosts on the network without manual intervention. For example, when an enterprise changes the Internet-connected ISP, a new global address prefix is obtained from the new ISP. The ISP uploads the address prefix from its router to the enterprise router. As the enterprise router periodically publishes a notice to all hosts in the local connection to broadcast the vro, all the hosts in the enterprise network will receive a new address prefix through the vro announcement. After that, they will automatically generate new IP addresses and overwrite the old ones.

Related Articles]

• IPv4 to Ipv6: Internet Development Trend
  

• Rip route configuration of ipv6
  

• Analysis and Comparison of IPv6 and NAT