The long and arduous way to IPv6


The argument that will be treated in this article can not be defined as “easy” or better “for everyone” since it provides the basics of  the binary code, but I will try to explain some concepts that sooner or later will be part of our lives in the near future.

We’ll start from the base… How is a computer identified on the internet (or LAN)?

The answer is: “Through an IP address“.

An IP address is a number that uniquely identifies, within a single network,  all devices connected to a computer network that uses standard Internet Protocol (IP). Each device (router, computer, network servers, printers, certain types of telephones…), therefore, has its address.

Simply, an IP address can be seen as the equivalent of a street address of devices connected on the internet.

Just as a street address identifies a building, so an IP address uniquely identifies a specific computer or any other network device or network. In turn, in a network can be used more valid IP addresses locally similar to the numbering of the interior of a building.

However, the ip address than the address road can ‘change very easily as a function of many variables. Can ‘be concealed or falsified, more or less legally.

Compared at road alone says nothing of who they are if not suppliers of internet access. Together with other data can ‘be used for profiling users.

Specifically, within a network at each interface connected to the physical network is assigned a unique address, so as to make possible the communication between a computer and the other.

These IP are formed in turn by a series of bits arranged in a specific sequence.

But how many IPs are there? How do they work?

For the answer to these questions we must know that the type of IP networks that are used in our current v4. But what does v4? It means that are composed of 8 bits arranged into 4 series.

Their structure determines their  maximum number around 4.5/5 billion addresses.

And, curious thing  to know that they are exhausted in February 2011.

But we let’s discover how we can remedy the problem.

With IP addresses (v4) now exhausted, the IS (Internet Society) is preparing the launch of IPv6, the successor of IPv4 Internet Protocol version to allow websites to continue their growth.

AT & T has recently indicated its deadline for starting is June 6, stressing out that only 1 % of its residential customers will begin to browse in “IPv6 mode.” As for mobile users, AT & T believes that they’ll need to wait much longer.


What is IPv6?

IPv6 is a new protocol that adds services simplifying configuration and network management. The current IP addresses are limited and are sold in the first two months of 2011. IPv4 could contain 32-bit addressing and managing approximately 4.5 / 5 billion addresses. With IPv6, however, can handle a much larger number and count on 128-bit reserved addresses.

For the industry experts the IPv6 is the  single most important update in the history of the Internet. Furthermore, content providers already decided to implement IPv6 and leave it on. John Curran, president and CEO of the American Registry for Internet Numbers, said:

“Going forward, Internet users will be forced to go through gateways to reach the transition to IPv6 companies updated with slower connections and services. The good news is that Internet users globally will not notice any significant difference in their daily lives.”


IPv6 packet

What is a IPv6 packet made of ? What is the difference between Unicast ,Multicast and Anycast  address type?

As it is defined in the document RFC 3513 , the structure addressing IPv6 provides three types of addresses:

  • Unicast: Identifies a single node
  • Multicast: identifies a group of nodes and the traffic destined to a multicast address is forwarded to all nodes in the group
  • Anycast: identifies a group of nodes and the traffic destined to an anycast address is routed to the nearest node of the group


IPv6 Unicast addresses

To keep the parallelism with IPv4, IPv6 addresses are segmented into a prefix and an interface identifier (ID interface), obtained in various ways: made of the physical interface (EUI-64), acquired through DHCPv6, set manually or generated cryptographic (RFC 3972).


Link Local addresses

When a node becomes operational, each interface obtains an address that can be used only for communication with other hosts on the same link. This address is used for communication only on that link and find operations such as neighbors or router.


Unique Local Unicast addresses

This address should be used only within the confines of a default domain: traffic that uses only local unicast addresses is not permitted beyond the limits of the domain. In this way, it becomes easy to connect separate local domains, since there are no collisions of addresses.


Unicast Global addresses

Unicast Global addresses are designed to be used on the IPv6 Internet. In order to facilitate its own aggregation, after numerous efforts to develop a flexible structure (RFC 2373), a strict policy allocation has been established  to maintain a simple format (RFC 3587).


IPv6 anycast addresses

As specified in RFC 3513, when the same unicast address is assigned to multiple nodes, it becomes an anycast address. A packet with an anycast destination address is forwarded to the nearest interface configured with it. Typically the anycast is implemented to replicate important network resources such as DNS, web servers, thus providing redundancy and load distribution.


IPv6 Multicast addresses

IPv6 multicast addresses replace the multicast broadcast addresses in control messages, then coating a critical role in the operations of the IPv6 network. A multicast address identifies a group of interfaces and can not in any way be used as the source address. IPv6 multicast addresses are distinguished by having the first 8 bits set to 1 (FF00 :: / 8).


What is expected of IPv6?

Do you think that IPv6 will solve the problems related to the use of the techniques of Network Address Translation ( NAT ) in use with IPv4? Do you think that IPv4 addresses are sufficient? The mechanisms of  address preservation could not cope with the demand for long global IP addresses. If conservative studies, based on growth rates of the past and present, estimated the exhaustion of IP addresses by February 2041, after more aggressive models foresaw dates much closer as early as 2009. The proliferation of communication gadgets, has been a stimulus in the use of the Internet that could not be inferred from the previous data.

Nell’IPv4 private addresses have always been associated with addresses that are not publicly registered. In an ideal world, privately addressed hosts are confined to a private network, but only hosts with public addresses are able to access the public domain, however, most of them must leave the confines private network at some point, but since there are not enough public addresses for all hosts on the private network, additional methods are needed for interfacing with the public domain.The simplest of all is the process of Network Address Translation ( NAT ).


IPv6 and NAT

The NAT has brought the best and the worst in the use of the IP protocol. Network Address Translation was born in the 90ies as a short-term solution, allowing address reuse and problem of their exhaustion. It worked well, with a critical  period in 1993, but over the last 10 years is less worrying.  The NAT has been widely used all over the Internet, going as justification, beyond the conservation of addresses from security to privacy by preventing redirection mechanisms to provide high availability.

Although NAT has damaged the use of many applications, the IPv6 offers solutions to all the problems that the NAT caused. The NAT  entails a series of problems, the most significant  is that it employs a model of asymmetrical connection. For its way of works, it is mainly used for client-server sessions, where a client with a private address connects to a server located on a public network. Anyone outside the private network would have trouble reaching a host behind NAT.


IPv6 and Voice

Another problem caused by NAT is that it does not look for any information beyond the IP, TCP, UDP and ICMP. Some applications, typically H.323 , have addresses springs hidden in their messages, which are not translated by NAT. Only address present in the IP header is modified. To resolve this problem, therefore, the complex application-level gateway is required.

IPv6, with its large address space (up to 128 bit), eliminates the need for NAT and at the same time provides functionality similar to the “perceived” benefits of NAT, such as multi-homing , privacy of the end-system, tracking the use of resources, darkening of the topology.



IPv6 has its roots in the fundamental concepts of IPv4 and draws on operational experience: the shortage of addresses gives us another opportunity to see how things can be improved, now that we know what to expect from the new protocol.

Matteo Indennimeo

*Editor’s note: the author was a Neomobile emloyee until 2013