Dual-stack networks

Dual-stack networks refer specifically to nodes that run IPv4 and IPv6 protocol stacks at the same time. However, other protocol stacks could also be included under the term in its original guise. In IPv4/IPv6 deployments these types of node can act as a protocol converter between the two networks and are sometimes referred to as IPv4/IPv6 nodes.

The IPv4 standard, which has served as the basis of the internet since it was first standardised in 1981, can no longer support the number of addresses that the ever-burgeoning internet requires. Availability of addresses is slated to run out between 2009 and 2012, according to many estimates. As a consequence, IPv4 is currently being replaced by IPv6 networks which can support a near infinite number of addresses. However, migration is not a switch off, switch on proposition and IPv4 networks will need to run alongside and interoperate with the replacement IPv6 infrastructure as it is deployed. It would simply be too costly and complex to attempt to affect such a rapid upgrade so IPv4 will need to run alongside IPv6 for many years to come.

The concept of dual-stack networks, with particular regard to IPv4/IPv6 migration, has been in existence since the need to migrate from IPv4 to IPv6 became apparent. On that level, it is an ongoing debate that has developed for most of this decade. However, in recent months several factors have come together to place greater emphasis on the subject. ICANN, the governing body that oversees internet access and protocols started to add IPv6 records to its root domain name servers this February, thereby opening the door to solely IPv6 networks. Prior to this, IPv6 users had to also retain IPv4 addresses in order to use domain names. That said, a great distance remains to be covered before the internet becomes an exclusively IPv6 environment.

The topic is coming to further prominence more recently with the announcement by the European Commission that it plans to force the pace of migration to IPv6. The Commission has set out a number of proposals to ensure that a quarter of European businesses, public sector bodies and households are using IPv6 by 2010. In addition, many carriers are well underway with their transitions – although, again, these are massive projects that will take years rather than months to come to fruition.

Aside from impractical and immensely costly total network upgrade, another method of achieving concurrently operational IPv4 and IPV6 networks is to use tunnels to carry one protocol inside another. In this case, a tunnel can take IPv6 packets and encapsulate them in IPv4 packets to be sent across areas within a network that haven't been upgraded to IPv6. Tunnels can be created where there are upgraded areas of IPv6 separated by unmodernised areas of IPv4, which is the norm during the early stages of this transformation. Tunnels can also be created in situations where the upgrade is further advanced and IPv4 packets need to be encapsulated in IPv6 packets to link up the now minority areas of IPv4 within a network.

There are two types of tunnels; manual and dynamic. Manually configured IPv6 tunnelling requires configuration at both ends of the tunnel, while dynamic tunnels are created automatically based on the packet destination and routing. Dynamic tunnelling techniques simplify maintenance but static tunnels make traffic information available for each endpoint, thereby providing extra security. Tunnels will require constant monitoring and alteration as the network transitions and, since they will need to be removed as the IPv6 deployment progresses, they will become just a transitional technique.

Dual-stack nodes can have three modes of operation. When the IPv4 stack is enabled and the IPv6 stack is disabled, the node behaves like an IPv4-only node. When the IPv6 stack is enabled and the IPv4 stack is disabled, the converse is true and it behaves like an IPv6-only node. Finally, when both the IPv4 and IPv6 nodes are enabled, the node can use both protocols.

Dual-stack devices must run both IPv4 and IPv6 at the same time and therefore need to save two sets of commands as well as compute, maintain, and store two sets of entries. In addition, a gateway device needs to convert packets between IPv4 and IPv6 and re-encapsulate the packets. Therefore, devices running two stacks assume greater loading than those running only one stack. As a consequence, there are greater requirements for device performance and more complicated maintenance and optimisation.