124 Part Two IPv6 Protocols
privacy, also makes IPv6 a more desirable protocol for commercial uses
that require special treatment of sensitive information or resources.
7.1 The IPv6 Address Space
The designers of IPv6 could have simply grafted a larger address space
onto the existing IPv4 addressing architecture—but doing so would cause
us to miss out on a huge opportunity for improving IP. Changing the
entire addressing architecture provides an incredible opportunity not only
for improving efficiency of address allocation but also for improving IP
routing performance.
The IPv4 address space was divided into several different classes based
on the values of their high-order bits. The IPv6 address space is also
divided into different categories based on high-order bits, as shown in
Table 7–1, although most of the address space is unassigned as of yet. The
global unicast address space takes up fully one-eighth of the entire address
space, with all global unicast addresses sharing the three high-order
bits 001. Other allocations are discussed later. All these allocations still
leave roughly 85% of the IPv6 address space unassigned, with no current
plans to assign them for now. At the same time, the allocations that have
been made should be more than ample for the foreseeable future.
As of 2003, the Regional Internet Registries (RIRs) are allocating /32 net-
work prefixes to organizations requesting them, and those allocations may
be expanded in the future if necessary. So far, there is no provision for
suballocations of network blocks smaller than /48, although there are
proposals on the table that would permit such “micro-allocations.”
In 1995, RFC 1884, “IP Version 6 Addressing Architecture,” allocated
a full quarter of the address space for two different types of unicast
addresses: one-eighthfor provider-based unicast addresses and one-eighth
for geographic-based unicast addresses. The intent was to offer addresses
that could be assigned based either on who provided network service to
the address holder or where the subscribing network was located. Provider-
based aggregation would have required networks to take on aggregatable
IP addresses based on the source of their Internet access. However, this
Chapter 7 IPv6 Protocol Basics 125
Allocation Prefix Fraction of
(binary) address space
Unassigned 0000 0000 1/256
Unassigned 0000 0001 1/256
Reserved for NSAP Allocation 0000 001 1/128
Unassigned 0000 01 1/64
Unassigned 0000 1 1/32
Unassigned 0001 1/16
Global Unicast 001 1/8
Unassigned 010 1/8
Unassigned 011 1/8
Unassigned 100 1/8
Unassigned 101 1/8
Unassigned 110 1/8
Unassigned 1110 1/16
Unassigned 1111 0 1/32
Unassigned 1111 10 1/64
Unassigned 1111 110 1/128
Unassigned 1111 1110 0 1/512
Link-Local Unicast Addresses 1111 1110 10 1/1024
Site-Local Unicast Addresses 1111 1110 11 1/1024
Multicast Addresses 1111 1111 1/256
Table 7–1: Initial IPv6 address space allocation map, from RFC 3513.
approach was seen to be less than a perfect solution for very large organi-
zations with far-flung branches, some of which would require service from
different providers. Provider-based aggregation would add even more IP
address management headaches for these large organizations.
Steve Deering proposed geographic-based allocation as an alternative in the
Simple Internet Protocol (SIP, a precursor to SIPP, see Chapter 4). These
addresses, unlike provider-based addresses, would be allocated on a per-
manent basis much as IPv4 addresses have been allocated. These addresses
would be based on geographic location, and providers would have to main-
tain additional routes to support these networks outside the aggregatable
portion of the IPv6 address space.

Get IPv6, 2nd Edition now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.