Inside IPv6

August 1, 2008


Think about how many devices we use in our personal and business lives that are connected to a network. When the Internet was first created, it was not intended for use by the general population. No one, including Al Gore, anticipated the Internet would find today's widespread use.

We are now running out of available IP addresses under the current addressing scheme called IPv4, which is limited to a maximum of 4,294,967,296 (232) addresses. This may seem like a lot of addresses, but consider that, according to the Internet Corporation for Assigned Names and Numbers (ICANN), the available IPv4 addresses will run out by 2011.

IPv6 Advantages
  • Easy configuration — Local devices will recognize each other and auto configure as necessary.
  • Ability to create ad-hoc networks quickly — good for rapidly connecting local devices and creating a temporary network
  • Supporting 3G and newer wireless telephone networks — Each telephone can have one or more dedicated IP address.
  • Will support larger packets up to 32Gb each — Yes this is correct you are currently limited to 64kb with IPv6.
  • Better security
  • Better quality of service

IPv4 vs. IPv6

There are many significant differences between the two addressing schemes, but the most obvious is that IPv6 was designed to handle about 3,403 × 1038 different addresses. That is approximately 1 trillion more than IPv4.

In terms of architecture, the two schemes are very different. IPv4 is categorized as classful, which essentially means that users requiring a static address were typically assigned a class B (65,534) or C (254) address. This is extremely inefficient, primarily because most assigned IP addresses were not needed. For example, if a company needs 300 IP addresses, it would waste 65,234 potential addresses. Now enter IPv6, which uses a scheme called Classless Inter-Domain Routing (CIDR), which permits the allocation of IP addresses to users based on the amount of addresses required. CIDR permits aggregation of contiguous addresses in a single supernet, which decreases the amount of data needed for routing tables or the complex sub-netting currently needed for some IPv4 configurations.

The key to CIDR is that all the routing information is included within the address using groups of CIDR blocks. It is prefix-based, which means that information about each packet, such as the address length, is specified in the header. For example, IPv4 information is included in the initial block, which ensures compatibility with IPv4 as well as IPv6 addressing.

The addressing will also look quite different. As a simple example, the well known private IP address (under IPv4) of 168.192.0.1, will be fe80::192.168.0.1. Several conversion calculators can be found on the Web.

IPv6 addresses are written in eight groups of four hexdecimal digits. One feature of this addressing scheme is that leading zeros can be omitted for each group, and if one group contains four zeros it can be replaced by double colons (::).

IPv4 reserves a block of IP addresses for local use. These addresses start with 192.168.xxx.xxx. IPv6 also provides private addresses, known as Unique Local Addresses (ULA), all of which begin with the prefix fc00::/7. In addition, IPv6 utilizes a pseudorandom algorithm that provides another layer of protection against the possibility that other network devices will end up with the same address and cause network collisions.



Changeover to IPv6

The transition to IPv6 will require hardware and software upgrades. Routers, network switches and many network interface devices may need to be replaced or upgraded. Fortunately, many manufacturers can provide the upgrades through firmware and/or software upgrades. Most of the new versions of operation systems, such as Vista, Mac and Linux, already have the IPv6 client. You can get the network clients for most older versions of these OS from the manufacturer. These network clients install and configure easily and, in most cases, are independent of the IPv4 client, so compatibility across a network should not be affected.

The real challenge will be to make sure your entire network is capable of passing both IPv4 and IPv6 packets. Any device not properly upgraded will likely prevent IPv6 packets to pass, which would tend to isolate IPv6 traffic to only a portion of your network. While it is possible (and probably necessary) for the two protocols to work in the same network, you will need to pay particular attention to how they will coexist in an on-demand/streaming environment such as that originating from digital audio/video servers and workstations.

To make the transition easier, the Network Working Group publishes a document called Basic Transition Mechanisms for IPv6 Hosts and Routers, or RFC 4213, that describes different methods to make the transition to IPv6. Essentially it describes three primary methods: Dual Stack, Tunneling and Translation.

Dual stack is simply a method where an additional software client supporting the IPv6 protocol is loaded to the device. This may be a software or firmware upgrade. You are familiar with this from loading different network clients on your PC. There might also be a new client written that supports both IPv4 and IPv6 in a single client.

Internet Resources

IP tunneling is a method that encapsulates a packet and allows it to be routed to a specific destination, essentially creating a dedicated pipeline or tunnel from two points. As it turns out, this is also a very efficient method to carry IPv6 packets over IPv4 networks. The IPv6 packet is encapsulated within an IPv4 packet at the entry point. When the packet reaches the destination, the IPv6 packet is stripped. The configuration information is maintained at the destination. This is called configured tunneling.

Translation is a method that uses an external black box to handle the exchange of packets.

The abundance of available addresses will finally make it possible to have direct access to individual devices tied to any network with relative ease. At the station level, implementing IPv6 should be relatively easy, providing you take the time to survey and evaluate each device connected to your network. Make sure you contact the manufacturer of any digital audio system or other IP-based equipment to find out what upgrades are required to support an IPv6 infrastructure. If your equipment is older and not going to be supported, don't worry. Realistically, you have some time before an upgrade to IPv6 will be needed, but make sure you provide for replacing that equipment in the next capital budget. If you are part of a larger group, the IT department will probably publish a timeline showing how the organization will migrate to IPv6, but it wouldn't hurt to ask the question if you haven't received any notices.


McNamara is president of Applied Wireless, Cape Coral, FL.


Comments