Something between IPv4 and IPv6 (1)

For IPv4, lack of IP addresses, network security vulnerabilities, bandwidth problems, and so on. Now the advent of IPv6 solves these problems, but for popularization, it has a considerable problem. To solve this problem, we need to do a good job of communication between IPv4 and IPv6.

(1) Dual Stack Model (RFC2893)

In this model, any node is fully dual-stack. In this case, communication between IPv4 and IPv6 does not occur. However, this mechanism assigns an IPv4 address to each IPv6 Site. This method cannot solve the problem of insufficient IPv4 address resources, and it is difficult to meet the requirement as IPv6 sites increase. Therefore, this method can only be used in the early change process.

(2) Limited Dual Stack Model (RFC2893)

In this model, the server and the router are still dual-stack, rather than the server host only needs to support IPv6. This mechanism can save a lot of IPv4 addresses, but communication between pure IPv6 and pure IPv4 nodes may cause problems. To solve this problem, it must be used in conjunction with other technologies.

(3) SIIT (Stateless IP/ICMP Translation, RFC2765)

SIIT defines the translation method between IPv4 and IPv6 packet headers. This translation is stateless, So translation is required for each group. This mechanism can be combined with other mechanisms (such as NAT-PT) for communication between pure IPv6 sites and pure IPv4 sites, however, this technology is unavailable in an environment that uses network-layer encryption and data integrity protection. The pure IPv6 node and the pure IPv4 node communicate with each other through a SIIT converter. The IPv6 node displays the host with a IPv4mapped address and uses an IPv4 translated address. If the destination address in the IP Group sent by the IPv6 host is a IPv4mapped address, the SIIT converter knows that the IP Group needs to undergo protocol conversion.

(4) NAT-PT (Network Address Translation-Protocol Translation, RFC2766)

NAT-PT is to do IPv4 and IPv6 address translation (NAT) at the same time in IPv4 and IPv6 packet header and semantic translation (PT ). It is applicable to communication between pure IPv4 sites and pure IPv6 sites. For high-level protocols with embedded address information such as FTP, The NAT-PT needs to work with the gateway at the application layer to complete the translation. The NAT-PT can be realized by using port information on the basis of the NAPT-PT, which has no essential difference with the NAPT under IPv4.

The principle of NAT-PT is similar to that of SIIT. The improvement is to apply the NAT under the traditional IPv4 to the selection of IPv4 addresses in SIIT. A major disadvantage of SIIT is that a large IPv4 address pool is required for dynamic allocation by IPv6 applications. This IPv4 address pool restricts SIIT applications to a large extent. The NAT-PT uses the traditional NAT technology under IPv4 to allocate IPv4 addresses, so that the IPv4 Address Allocation pool can be formed with a few IPv4 addresses, A protocol conversion service can be used for a large number of applications that require address translation.

In terms of implementation, without the support of the DNS-ALG, only the communication between IPv6 and IPv4 can be realized, on the contrary, the package will be discarded. If DNS-ALG is supported, two-way communication can be realized. Some applications require a certain degree of address stability, and the NAT-PT can be configured to provide static ing from V6 to a specific V4 address.

This mechanism is suitable for the initial phase of transition, so that hosts based on the dual-protocol stack can run IPv4 applications and IPv6 applications to communicate with each other. This technology allows IPv6-only applications that do not support IPv6. This mechanism requires that the host must be dual-stack. At the same time, three special extension modules should be inserted in the Protocol Stack: domain name parser, address er, and translator, NAT-PT is used in the protocol stack of the host.

(6) BIA (Bump-In-the-API, Internet Draft)

Similar to BIS, this technology only translates packets at the API layer rather than the Protocol Stack layer. Therefore, its implementation is simpler than BIS, because you do not need to translate IP headers. The main difference between BIS and BIA is that BIS is used in systems without IPv6 protocol stacks, while BIA is used in systems with IPv6 protocol stacks.