IPv4-to-IPv6 Transition Mechanisms

Implementing IPv6 requires a transition mechanism to allow interoperability between IPv6 nodes (both routers and hosts ) and IPv4 nodes. The transition mechanism is the key factor in the successful deployment of IPv6. Because millions of IPv4 nodes already exist, upgrading every node to IPv6 at the same time is not feasible . As a result, the transition from IPv4 to IPv6 happens gradually, allowing nodes to be upgraded independently and without disruption to other nodes. While a gradual upgrade occurs, compatibility between IPv6 and IPv4 nodes becomes a requirement. Otherwise, an IPv6 node would not be able to communicate with an IPv4 node.

Transition mechanisms allow IPv6 and IPv4 nodes to coexist together in the same network and makes gradual upgrading possible. Two transition mechanisms are supported by the JUNOS software:

• Dual IP layer operation ”Allows a node to support both IPv6 and IPv4 Internet protocols. This mechanism is also known as dual stack.

• Configured tunnels ”Allows IPv6 packets to be encapsulated into IPv4 headers and sent across an IPv4 infrastructure.

Dual IP Layer

Dual IP layer is a transition mechanism in which a node (router or host) supports both IPv6 and IPv4 simultaneously . Dual IP layer support indicates a node's capability to send and receive both IPv6 and IPv4 packets. This transition mechanism involves no data encapsulation (such as tunneling). A node that has dual IP layer capabilities can interoperate with IPv6 nodes and IPv4 nodes directly.

A node that supports dual IP layer can disable one of the IP layers . If a dual IP layer node disables the IPv6 layer, the node acts as an IPv4 node only. If that node disables the IPv4 layer, that node acts as an IPv6 node only.

You can configure a dual IP layer node with either an IPv6 address or an IPv4 address. A dual IP layer node can acquire its IPv6 address through an IPv6 mechanism (such as stateless address autoconfiguration) and its IPv4 address through an IPv4 mechanism (such as Dynamic Host Control Protocol [DHCP]).

The Domain Naming System (DNS) uses resource records and resolver libraries to map between IP addresses and hostnames. IPv4 uses record type A, whereas IPv6 uses record type AAAA. Dual IP layer nodes provide resolver libraries that can handle the new record types.

When a query locates an AAAA record holding an IPv6 address and an A record holding an IPv4 address, the resolver library sends back one of the following types of address to the application making the query: IPv4 address, IPv6 address, or both. If the library sends back an IPv4 address, the application making the query uses IPv4 to communicate with the node. If the library sends back an IPv6 address, the application making the query uses IPv6 to communicate with the node. If the library sends back both IPv6 and IPv4 addresses, the application must choose based on ordering of the addresses sent by the resolver library. If an IPv6 address is the first in the order, the application chooses IPv6. If an IPv4 address is first in the order, the application chooses IPv4.

Configured Tunnels

IPv6-over-IPv4 tunnels are defined in RFC 2893, Transition Mechanisms for IPv6 Hosts and Routers .

If a Tunnel Services PIC is installed in the router, you can configure IPv6 over IPv4 tunnels. The JUNOS Internet software supports the IP over IP (IP-IP) and Generic Route Encapsulation (GRE) tunnel encapsulations . A configured tunnel is a point-to-point connection across an existing IPv4 network infrastructure. IPv6 packets are encapsulated in IPv4 headers and sent across the IPv4 infrastructure through the configured tunnel. You manually configure configured tunnels on each end point. A configured tunnel cannot go through a network ad-dress translation (NAT) at any point along the way to the destination.

To configure a tunnel using IP-IP encapsulation, include the following statements:

 [edit interfaces]  ip-  fpc  /  pic  /  port  {   unit  logical-unit-number  {     tunnel {       source  address  ;       destination  address  ;     }     family inet6 {       address  address;  }   } } 

To configure a tunnel using GRE encapsulation, include the following statements:

 [edit interfaces]  gr-  fpc  /  pic  /  port  {   unit  logical-unit-number  {     tunnel {       source  address  ;       destination  address  ;     }     family inet6 {       address  address;  }   } } 

You can configure multiple logical units for each interface. You can configure only one tunnel per unit. Each tunnel interface must be a point-to-point interface. Point to point is the default interface connection type, so you do not need to include the point-to-point statement at the [edit interfaces interface- name unit logical-unit-number ] hierarchy level. You must specify the configured tunnel's destination and source addresses. The remaining statements are optional. You must configure inet6 address family on the logical interface to create a configured tunnel. Configured tunnels are bidirectional.