Previously commsafrica.com took a look at the efforts behind the IPv6 drive and some of the reasons for implementing a new IP protocol. This week we look at the addressing of the protocol underlying this work in progress.
Designed as a successor to IPv4 the next generation IP V6 has Extended Routing and Addressing Capabilities. This means that the addresses are 128 bits long with identifiers for individual interfaces. The difference at this level is that the addresses are assigned to interfaces and not nodes. . As interfaces belong to a single node the interface`s unicast addresses can be used as an identifier for the node, this means that multiple IPv6 addresses may be assigned to a single interface.
Addresses
Three address types of IPv6 may be used to identify a single interface. These are the multicast, anycast and unicast. Unicast is used for a single interface, anycast will identify a set of interfaces where one member of a set will receive the packet. Multicast will deliver the packet to all of the interfaces identified in the group.
IPv6 supports addresses four times the number of bits of IPv4. This translates to 4 billion times the size of address space.
For the sake of practicality it was necessary to create a hierarchy that would be able to deal with the assignment and routing of addresses to enhance the efficiency of the space used. Taking into account the present telephone systems and the current settings of the Internet and the IEEE 802 nodes it was calculated that this new addressing system would provide for 1564 addresses for each square metre of the planet Earth.
The leading bits, called the Format Prefix, specify the type of address. The allocation supports provider addresses, local use addresses and the multicast addresses. Taking the future into account space has been allocated to the IPX addresses, interconnect addresses and NSAP with some space undefined and unreserved.
Unicast
IPv6 makes provision for several types of unicast addressing. The first of these is the provider unicast address. These are the addresses used for global communication and similar to the addressing of IPv4. In this instance the first three bits will identify the address as a provider unicast address. The following field will identify the Internet address registry, which in turn will identify the next field that will identify the Internet service provider. The segment after that will identify the subscriber, this in turn leads to the subscriber identification if there are multiple users on the address. The subnet and interface identification will indicate the specific subnet on the physical link and the interface among the group as shown by the subnet identification.
To accommodate the subnet or local use addresses as a unicast address system there are two types defined as link-local and site-local. The site-local is designed for use on a single site while link-local is on a single link. Both these addressing systems utilise the interface identifier that must be unique in the domain. Most instances the IEEE 802 address will be used. The subnet identifier will indicate a specific subnet of the site. The combination of subnet and interface identifiers enables a large private Internet to be built without other address allocation.
For companies that are as not yet connected to the Internet local-use addresses will allow for addressing without the need for an address prefix from global Internet space. Should the company have plans to gain a presence on the Internet then the subnet and interface identification may be used in together with a global prefix
Anycast
The IPv6 anycast address is the address assigned to more than one interface on different nodes. Ideally a packet sent to an anycast address is directed to the nearest interface in accordance with the protocol`s determination of distance.
The anycast address will allow a node to choose an Internet service provider it wants to carry the traffic. This would be achieved by configuring the anycast to select certain routers from a set of pre-identified routers of the various service providers. The anycast address may be used as an intermediate means of addressing and delivery as a packet gets routed through various routers on its way to a final destination.
Anycast addresses share the same space as that allocated to the unicast addressing, this makes the anycast addresses indistinguishable from unicast based on the syntax of the addressing.
Multicast
The multicast address uses the identifier for a group of interfaces, these in turn belong to multicast groups. The first eight bits at the header of the packet will identify that the address is a multicast address. The following four bits are reserved for the `flags`, which will indicate that the multicast is a permanent group with a global address or a temporary group. The four bits following these are used to define the scope of the group and may range from a reserved organisation scope to node-local, link-local and site local.
The IPv6 transition mechanisms also make provision for the addressing capabilities of IPv4, which will be embedded into the next generation protocol. The provision made for addressing in the IPv6 should go a long way to making the whole implementation slightly easier with the added ability of being ale to grow.
Share
