IT Management

Mar 1 ’03

It may be new to your organization, but Transmission Control Protocol/Internet Protocol (TCP/IP) has been around quite awhile. We’re currently using what’s known as IP version 4 (TCP/IP V4). Each version has represented significant change to meet the demands applications have placed on our IP networks. Today’s new interest in this protocol by businesses and the impact of the Internet are forcing another new wave of changes to the IP protocol — resulting in IP V6 (also called IPNG). The Internet Society, a global group that sets architecture standards and watches out for the health and well-being of IP networks, defined the IP V6 standard.

Key features include:

  • Support for millions of potential users and devices with a larger address space
  • Class of service, which allows widespread use of time-dependent applications such as streaming audio and video
  • Built-in security, which provides organizations and individuals privacy as activities now traverse the Internet and other IP networks
  • Multicast and encapsulation techniques that reduce traffic placed on global networks
  • Reduced IP host configuration.

Since products from multiple vendors will begin shipping in 1997 with IP V6 support, let’s look at each of these features and what their impact will be on your IP networks.  

Addressing

The trade press has been discussing the limitation of the existing IP V4 address space for the past year. Each device in an IP network has a unique global IP address that lets applications communicate and lets each of us work more effectively. In an IP V4 network, a 32-bit dotted decimal numbering system was used. This allowed for the definition of just over 4 billion addresses. The easiest way to support more devices is to increase the number of spaces used to represent an address. The standard for IP V6 supports a 128-bit hexadecimal colon numbering system (see 1).

First, let’s review the existing IP V6 address. The dotted decimal system indicates that only the numbers 0-9 will be used to represent the addresses. The 32 bits are split into four 8-bit (4 bytes) sections separated by a decimal point. This results in addresses such as 192.84.3.1 and 9.128.35.128.

IP V6 uses the hexadecimal numbering scheme 0-F and breaks the address into 16 8-bit (16 bytes) sections separated by colons. With the 128-bit address space, IP V6 uses a more structured hierarchical system to help reduce the traffic used to communicate. This new system will allow traffic to be easily directed to the destination with a more direct path. It will work much like our existing ZIP code system, where ZIP codes that start with 0, 1, 2 are on the East Coast and those that start with 7, 8, 9 are on the West Coast. How this hierarchy is implemented will depend on Internet Service Providers (ISPs), the Internet Society, and major corporations. The global distribution of addresses for IP V6 will continue through the same organizations providing this support for IP V4. In the U.S., the contact is the InterNIC at ‘ftp://rs.internic.net/templates/ internet-number-template.txt’.

You may be concerned that having each user of an IP host key in this 16- byte address opens the way for a tremendous amount of errors. In a LAN environment, the addresses scheme plans on using the 48-bit (6 bytes) LAN address plus an 80-bit (10 bytes) header. The 80-bit header will be provided to the IP host by the router or other management device on the network. Other detailed schemes are being defined for Internet access through an ISP. This relieves the user of any address configuration.

To support interoperability with IP V4, two specific variations of the IP V6 address are defined. An IP V4-compatible address can be converted to and from the older IP V4 address format. These are used when IP V6 systems need to communicate with each other, yet find themselves separated by an IP V4 network. Routers at the boundary of the IP V4 network can convert these addresses to IP V4 so they can transverse this older network.

You form an IP V4-compatible address by preceding 96 bits of zero to a legitimate 32-bit IP V4 address (see Figure 2). Using this address for compatibility mode defeats some of the advanced hierarchical schemes possible with an IP V6 network address. Asecond type of address, called IP V4-mapped address, is also defined. An IP V4-mapped address indicates a system that doesn’t support IP V6. As long as there are routers or other devices between the IP V6 and IP V4 device that can convert the addresses between the two addressing formats, the IP V4 and IP V6 device can communicate (see Figure 2). The definition of both IP V4-compatible and IP V4-mapped addresses were chosen to support the various checksum algorithms that many IP protocols use.

3 Pages