Although the IPv6 protocol offers a host of technological advances and innovations, its use must still be deployed by information technology (IT) staff in end-user organizations and ISPs. The deployment of IPv6 involves the planning and design of coexistence and migration strategies and the installation and maintenance of hardware and software. The resulting combination of IT staff, hardware and software resources, and time required for the transition makes the decision to deploy IPv6 a significant one, especially in light of other technology initiatives that may have higher visibility or better short-term benefits.
One must consider, however, that the Internet, once a pseudo-private network connecting educational institutions and United States government agencies, has become an indispensable worldwide communications medium that is an integral part of the increased efficiency and productivity for commercial organizations and individuals, and a major component of the world's economic engine. Its growth must continue.
To continue the growth of the Internet, IPv4 must eventually be replaced. The sooner IPv4 is replaced, the sooner the benefits of its replacement protocol are realized. The following sections present the key evidence in the case to deploy IPv6.
First, and most visibly, IPv6 solves the IPv4 public address depletion problem by providing an address space to last well into the 21st century and, probably, beyond. By moving to IPv6, mobile cell phones, personal data assistants (PDAs), automobiles, appliances, and even people can be assigned multiple public addresses. The growth of the devices connected to the Internet can proceed without restraint.
The Internet was principally a creation of educational institutions and government agencies of the United States of America. In the early days of the Internet, connected sites in the United States received IP network IDs without regard to summarizability or need. The historical result of this address allocation practice is that the United States has a disproportionate number of public IP addresses. For example, there are educational institutions in the United States that have more address space than the entire nation of the People's Republic of China.
With IPv6, public address prefixes are assigned to regional Internet registries, which, in turn, assign address prefixes to other ISPs and organizations based on justified need. This new address allocation practice ensures that addresses will be distributed globally based on regional connectivity needs, rather than historical origin. This makes the Internet more of a truly global resource, rather than a United States-centric one.
With IPv6, NATs are no longer necessary to conserve public address space and the problems associated with mapping addresses and ports disappears for developers of applications and gateways. More importantly, end-to-end communication is restored between hosts on the Internet by using addresses that do not change in transit. This functional restoration has immense value when one considers the emergence of peer-to-peer telephony, video, and other real-time collaboration technologies for personal communications, and that the next wave of devices that are connected to the Internet include many types of peer-to-peer devices, such as mobile phones.
Unlike IPv4 addresses, IPv6 addresses have a scope, or a defined area of the network over which they are unique and relevant. For example, IPv6 has a global address that is equivalent to the IPv4 public address and a site-local address that is equivalent to the IPv4 private address. Typical IPv4 routers do not distinguish a public address from a private address and will forward a privately addressed packet on the Internet. An IPv6 router, on the other hand, is aware of the scope of IPv6 addresses and will never forward a packet over an interface that does not have the correct scope.
There are many types of IPv6 addresses with different scopes. When multiple IPv6 addresses are returned in a DNS name query, the sending node must be able to distinguish their types and, when initiating communication, use a pair (source address and destination address) that is matched in scope and that is the most appropriate pair to use. For example, for a source and a destination that have been assigned both global (public) and site-local (private) addresses, a sending IPv6 host would never use a global destination with a site-local source. IPv6 sending hosts include the address selection logic that is needed to decide which pair of addresses to use in communication. Moreover, the address selection rules are configurable. This allows you to configure multiple addressing infrastructures within an organization. Regardless of how many types of addressing infrastructures are in place, the sending host always chooses the "best" set of addresses. In comparison, IPv4 nodes have no awareness of address types and can send traffic to a public address from a private address.
IPv6 is a streamlined version of IPv4. Excluding non-default QoS traffic, IPv6 has fewer fields to process and fewer decisions to make in forwarding an IPv6 packet. Additionally, the hierarchical and summarizable addressing structure of IPv6 global addresses means that there are many fewer routes to analyze in the routing table of organization and Internet backbone routers. The outcome is traffic that can be forwarded at higher data rates, resulting in higher performance for tomorrow's high bandwidth applications that utilize multiple data types.
IPv6 has been designed to support security and mobility as built-in features. Although one could argue that these features are available for IPv4, they are available on IPv4 as extensions, and therefore have architectural or connectivity limitations that might not have been present if they had been part of the original IPv4 design. It is always better to build features in rather than bolt them on.
The result of building in security and mobility to IPv6 is an implementation that is a defined standard, has fewer limitations, and is more robust and scalable to handle the current and future communication needs of the Internet.