In 1994 work began on a replacement for IPv4, which eventually became known as IPv6, for which the first standard was introduced in 1998. IPv6 differs from IPv4 in many ways, but the most important feature was the use of a 128 bit address structure rather than the 32 bit address structure utilized by IPv4. This provides for Trillions of addresses that should be sufficient to service current and future needs. Version 4 addresses are normally written in dotted decimal notation, but Version 6 addresses are normally seen in Hexadecimal notation.
Version 6 of the Internet Protocol is a completely new protocol, and although it can operate alongside Version 4, it is not compatible without some special configuration on network devices or translation. There is no short term plan to retire IPv4, instead there are a number of strategies that enable a migration from an IPv4 network to a dual IPv4 / IPv6 network. Addresses can be manually or dynamically assigned, with the four recognised ways to assign Interface Identifiers being:
Static Manual Assignment
Static Assignment using the EUI-64 interface ID
The largest proportion of the IPv6 address space is allocated to Global Internet addresses, of which some has been designated for multicast operation. However, there are other address types that can be used:
The Link Local address is a new concept, and they only refer to a local physical link. Routers will not forward packets using link-local addresses, they are only for local communication on a particular physical network segment.
Link Local addresses - Have a limited scope and are created automatically on an interface, and are only valid on a single physical link and can be used for Automatic Address Allocation, Neighbor and Router Discovery. When communicating with a Link Local address, the outgoing interface must be specified because every interface is connected using a 10 bit prefix beginning FE80.
Site Local or Unique Local Addresses are the equivalent of private addresses used with Internet Protocol Version 4. They originally used the prefix FEC0 but were redefined in 2005 to FD00 and another allocation with a prefix of FC00 has also been defined.
Some of the main features of IP Next Generation include:
A simplified header format that ensures that overhead is minimized and routers can process packets quickly and efficiently.
IPSec is built in to the protocol and so becomes mandatory, allowing the use of Authentication, Encryption and Key Exchanges necessary for VPNs. Extension headers can specifiy the particular security protocols in use.
Flow Labels to allow streams to be identified and prioritized.
Both Stateful and Stateless address configuration, so network administrators can choose DHCP or rely on Stateless Auto-configuration of addresses or even manually configure addresses.
There are no broadcasts with IPv6, but local neighbours can discover each other through the Neighbor Discovery Protocol.
Flexibility has been built into IPv6 through the use of extension headers to describe any additional features or demands not identified in the standard header. This allows new extension headers to be defined in future.
Although Broadcasting has been eradicated through the use of IPv6, there is a new concept known as anycast, where multiple devices running identical services can use the same anycast address and routers will route traffic to the nearest device running the same address. The term One to the Nearest is often used to describe anycasting.
Here at NSTUK we run a number of courses running IPv6 topics, including:
IPv6 Overview (1 Day)
Practical Introduction to IPv6 (2 Days)
Cisco Networking Part 1 (5 Days)
Cisco Networking Part 2 (5 Days)