IPv6 Addressing | Solaris 9 Network Administrator Exam Cram 2 (Exam CX-310-044)

IPv6 addresses are 128 bits long and, like IPv4, use three types of addresses, but they are not the same three types. IPv6 uses the following basic address types:

Unicast ” A unique address is assigned to a network interface, which communicates with a single host on the network.
Multicast ” A unique address is assigned to a group of systems. A message is sent from one host to all hosts that have the address assigned.
Anycast ” A unique address is assigned to a group of systems, but a message sent from one host is delivered to only the nearest host that has the address assigned, unlike the multicast address type, in which all hosts that have the address receive the message.

IPv6 does not use the broadcast address type as IPv4 does. With IPv6, a physical network interface can be assigned several types of IPv6 addresses. This is an important difference from IPv4. Remember also that IPv6 has introduced the anycast address type, which is not present in IPv4.

How IPv6 Addresses Are Represented

Each IPv6 address is divided into 8 octets, each consisting of 16 bits. An IPv6 address is written in hexadecimal, with each octet being separated by the colon (:) character. The general format is

 nnnn:nnnn:nnnn:nnnn:nnnn:nnnn:nnnn:nnnn

where each nnnn group of characters represents four hexadecimal digits (16 bits). An example of an IPv6 address is

 fe80:0000:0000:0000:0a00:20ff:feb3:4153

An IPv6 address can be written a number of ways because consecutive zeros can be shortened . The preceding address can be written in these ways:

The standard way of representing an IPv6 address:
```
 fe80:0000:0000:0000:0a00:20ff:feb3:4153 
```
Groups of 16-bits consisting of all zeros can be shortened to a single zero:
```
 fe80:0:0:0:0a00:20ff:feb3:4153 
```
Groups of 16 bits consisting of all zeros can be compressed using a double colon (::) character:
```
 fe80::0a00:20ff:feb3:4153 
```

The double colon (::) character that is used to compress 16-bit groups consisting of all zeros can be used only once in an IPv6 address. It is quite common to get a question on the exam that provides a number of IPv6 addresses and asks which ones are valid. For example, you might see something like the following IPv6 address:

 fe80::0a::20ff:feb3:4153

This address is invalid because it uses the double colon (::) character twice in the address representation.

Format Prefixes (FP)

A format prefix at the beginning of each IPv6 address specifies the type of IPv6 address that is being used. Table 7.1 identifies some of the common format prefixes in use.

Table 7.1. Common IPv6 Format Prefixes

Format Prefix (Binary)	Format Prefix (Hexadecimal)	Address Type
0000 0000		Reserved
1111 1111	FF	Multicast
001	2	Aggregatable global unicast
1111 1110 10	FE8	Link-local unicast
1111 1110 11	FEC	Site-local unicast

Notice that the anycast address type is not included in Table 7.1. Anycast addresses are allocated from the unicast address space.

You might also see an IPv6 address written as follows :

 fe80:0000:0000:0000:0a00:20ff:feb3:4153/10

The /10 at the end of the address signifies that 10 bits have been used to identify the address type. This is the format prefix, also known as the IPv6 subnet prefix. As you can see from Table 7.1, the address starts with fe8 and is a unicast link-local address. When an address is specified with the format prefix attached, it is known as prefix notation .

Unicast Addresses

The unicast address type is used to send IP datagrams to a single host, or interface. A number of unicast address types that each serve a particular purpose are described in the following sections.

Link-Local

The link-local address is intended to be used only on local network links; that is, a system using this kind of address can send IPv6 datagrams only to another system on the same physical network. A router will not forward this type of packet onto another network.

Site-Local

The site-local address is very similar to the link-local address, but can be forwarded (routed) through an internal intranet. The limit for this kind of address is the site rather than the same physical network.

Aggregatable Global

The aggregatable global address type is used for global communications, such as across the Internet. Aggregatable global addresses contain a hierarchical structure of identifiers, as follows:

The format prefix ” A 3-bit field that identifies the address type as aggregatable global and is set to 001 hexadecimal.
A top-level aggregator (TLA) ” A 13-bit number that identifies the Internet authority that assigned the next -level aggregator ”the Internet Assigned Numbers Authority (IANA), for example.
A next-level aggregator (NLA) ” A 32-bit number that is assigned to an organization, normally by an Internet service provider (ISP).
A site-level aggregator (SLA) ” A 16-bit locally managed field that is used by an organization to create its own addressing and subnetworking hierarchy. This field is similar to the subnetwork mask in IPv4, except that the number of subnetworks is much greater ”this field supports up to 65535 individual subnetworks.
An interface identifier ” The field that identifies individual interfaces on a network. This number must be unique and is normally based on the ethernet address associated with the particular interface. (The interface identifier calculation is discussed later in this chapter in the section "IPv6 Autoconfiguration.")

On the exam, if you have to enter any of the unicast address types, remember that link-local and site-local are hyphenated. You could possibly lose a mark if you forget the hyphen. Also, aggregatable global is not hyphenated.

Embedded IPv4 Address

The transition to IPv6 is going to take many years , so a mechanism was necessary to allow IPv6 and IPv4 to work together while the migration to IPv6 carries on. IPv6 systems that use this mechanism use a special IPv6 unicast address, called an IPv4-mapped IPv6 address , that can accommodate an IPv4 address in the lower 32 bits of the address. This address is used to represent an IPv4 address within an IPv6 address space. The general format of an IPv4-mapped IPv6 address is

 0000:0000:0000:0000:0000:FFFF:X.X.X.X

where

The first 80 bits are zeros.
FFFF is the flag that indicates that an embedded IPv4 address follows.
X.X.X.X is the 32-bit IPv4 address, represented in decimal dot notation, as it is in IPv4.

As an example, the embedded IPv4 address 192.168.28.28 would appear as the following IPv6 address:

 0000:0000:0000:0000:0000:FFFF:192.168.28.28

Additionally, an IPv4 “compatible-IPv6 address consists of the first 96 bits being all zero, followed by the 32-bit IPv4 address, again in decimal dot notation. This address is used by hosts and routers to transfer IPv6 packets on an IPv4 infrastructure.

Unspecified Address Type

The unspecified address type is used when the sending system does not have an IP address assigned, such as when a diskless client starts up. The address consists of all zeros and can be represented as follows:

0000:0000:0000:0000:0000:0000:0000:0000 ” Full representation.
0:0:0:0:0:0:0:0 ” In a shortened zero format.
:: ” In a compressed format.

Loopback Address Type

An IPv6 system, like an IPv4 system, uses a loopback address to send datagrams to itself. The IPv6 loopback address is shown in the three representations as follows:

0000:0000:0000:0000:0000:0000:0000:0001 ” Full representation.
0:0:0:0:0:0:0:1 ” In a shortened zero format.
::1 ” In a compressed format.

This is equivalent to the IPv4 address 127.0.0.1 .

Multicast Addresses

The multicast address is used to address a number of hosts simultaneously . All hosts (or interfaces) that belong to the same multicast group receive the message.

Multicast addresses include a 4-bit field, called scope bits , that determines how far a multicast IPv6 datagram is routed. The contents of the first octet of the IPv6 multicast address determines the "scope" of the datagram. The format prefix for multicast addresses is FF . The next 4-bit field is used for flags that are normally set to zero. The final 4-bit field is the scope where:

1 indicates a node-local address ” that is, the address is on the same host or system as the sender.
2 indicates a link-local address ” that is, the destinations are all on the same network link as the sender.
5 indicates a site-local address ” that is, the destinations are all on the same site as the sender.
E indicates a global address ” that is, the destinations are all on the Internet.

As an example, the IPv6 multicast address to send a datagram to all NTP servers located on the same site is

 FF05:0:0:0:0:0:0:101

Anycast Addresses

The anycast address is an IPv6 address that is assigned to more than one interface. A packet sent to an anycast address is received only by the nearest interface that has the address assigned to it ”nearest being interpreted by the routing protocol being used. (RIP, for example, would select the interface that can be reached in the fewest hops.)

Anycast IPv6 addresses are assigned from the unicast address space, so as soon as you assign a unicast address to more than one interface, it automatically becomes an anycast address.