understand how they are used and how to ease the transition from IPv4 to IPv6. Part one of this two-part series on Internet Protocol Version 6 (IPv6) addressing basics described the IPv6 address types and formats and the qualifiers defining the area of the network on which the address is valid. Part two describes how these address formats are used in an IPv6 network and how to convert an IPv4 address to IPv6. For companies that have decided to stay on IPv4, view this tip on the pros and cons of disabling IPv6 in Windows Vista.
Don't miss both installments of this two-part series on IPv6 addressing:
- IPv6 address types
- Transition IPv4 to IPv6: How address formats are used to convert IPv4
On initial startup, an IPv6 node automatically creates a link local unicast address. The link local address will be used to gather additional information needed to construct a global unicast address which will then be used to communicate across the Internet. The first step is to construct an Interface Identifier (Interface ID). The Interface ID is a 64 bit number created from the 48 bit hardware MAC address. An additional 16 bits must be added to the MAC address to fill out the required 64 bits. The same 16 bit quantity is always used; hex FF-FE is inserted between the first three and last three octets of the MAC address. After constructing the Interface ID, the node then combines the Interface ID with the prefix specifying link local unicast addresses to create a link local address that it can be used to communicate with other nodes on its link.
After constructing a link local address, the node can listen for a router advertisement message. Routers periodically send advertisements that carry information needed to construct a global unicast address. Instead of waiting for an advertisement, the node can send a router solicitation to the all-routers multicast address. Routers on the link reply with the same information that would be contained in a router advertisement. The information includes the prefix containing the high order network address bits specifying this particular site. After receiving information from a router, the node combines it with its Interface ID to construct a unique global unicast address.
Network renumbering on an IPv4 network usually requires the network to be shut down for a time to reconfigure the DHCP servers and gateways and make sure that all nodes have picked up new addresses from the DHPC servers. In order to map IPv4 to IPv6 addresses, IPv6 enables network renumbering without a need to shut down the network.
Network address prefixes are valid only for a limited period, but a router can send a new prefix in an advertisement at any time prior to the expiration of a previously sent prefix. When a node receives a new prefix, it will switch over to using the new prefix for new connections, but the old prefix will remain valid until its period of validity expires. As connections are terminated, use of the old prefix will diminish and eventually end, but there will be no need to shut down the network.
In IPv4, CIDR enables Internet backbone routers to aggregate routes to a limited extent. The designers of IPv6 recognized that the much larger IPv6 address would enable route aggregation to a greater extent by enabling service providers to assign addressees in a hierarchical manner. Earlier IPv6 specifications described a fixed two level hierarchy with a 13 bit Top Level Aggregation Identifier, a 24 bit Next Level Aggregation Identifier, and 16 bit Site Level Aggregation Identifier. More recently, recognizing that little is known at this point about how IPv6 networks will evolve, IPv6 designers removed the fixed fields from the specification. Each of the registries and backbone service providers can allocate fields within its assigned address space in an optimal manner based on the number and size of its subordinate registries and service providers to which it assigns addresses.
IPv4 to IPv6 transition
IPv6 provides the IPv4 Mapped IPv6 Address format to enable the transition from IPv4 to IPv6. This format consists of 80 zero bits followed by 16 ones followed by a 32 bit IPv4 address. A dual stack node will use the entire 128 bit address when sending to an IPv6 node and will use only the low order 32 bits when sending to a node supporting only IPv4.
IPv6 migration will be take many years, and require much relearning throughout the Internet community, but IPv6 designers have carefully created address formats that support facilities to ease network management and to enable efficient Internet operation for years to come. You can view the road to the new Internet Protocol in this IPv6 timeline for more information.
About the author:
David B. Jacobs has more than twenty years of networking industry experience. He has managed leading-edge software development projects and consulted to Fortune 500 companies as well as software start-ups.
This was first published in April 2005