IPv6: How It Works - and Will It Ever Be Widely Adopted?
In development since the mid-1990s, IPv6 (Internet Protocol version 6) has not yet been widely adopted, due in large part to compatibility difficulties with IPv4, which is the basis of virtually all Internet communications today.
IPv6’s importance rests on the fact that it is the only alternative to IPv4, at a time when IPv4 address space is becoming harder to obtain, and when major new deployments of IP technology, in developing countries and in handheld mobile devices, are being planned.
A new BuddeComm biennial report in Handbook format - 2008 Technology - Internet - Volume 4 - IPv6 - introduces managers and technical specialists to the technologies and controversies of IPv6. The BuddeComm Handbooks provide an overview and functional understanding of important technologies which are otherwise hard to grasp without one-on-one training or lengthy study of voluminous engineering-oriented material.
While it is widely assumed that IPv6 will one day be so widely adopted as to overtake and replace the current IPv4-based Internet, we argue that such ubiquitous adoption is not assured in the foreseeable future. At a time when many organisations are tempted to adopt IPv6, or are being pressured to do so by governments, this Handbook provides an alternative perspective to much of what has been written about IPv6 by its proponents.
The Handbook provides a comprehensive introduction to IPv6’s addressing system and to its commonality and differences with IPv4. We offer critical perspectives on some of IPv6’s purported benefits, such as its long 128 bit address system and the yet-to-be completed SHIM6 approach to multihoming. Multihoming is the ability to reliably connect to the Internet via two or more ISPs. SHIM6 is intended to enable this without the use of BPG-managed PI (Provider Independent) address space. Current BGP-based multihoming practice drives up costs of core routers for every ISP and is the primary cause of the Internet’s routing and addressing crisis.
IPv6 addressing is more complex and flexible than that of IPv4. We discuss the auto-configuration processes by which an IPv6 host computer develops its own IP addresses, with separate addresses for local and global use. We discuss Unique Local Scope addresses which enable a site to retain a fixed set of addresses, which are not globally routable, no matter which ISP they connect to the Net with.
IPv6 provides extensive support for automated assignment of IP addresses to hosts and routers, which is intended to make it more practical to renumber an entire network from one range of PA (Provider Aggregatable) address space to another, when changing ISPs. However, there are many challenges to achieving this securely and robustly, and many other aspects of networks beyond the reach of this system would require manual reconfiguration if a new address range was adopted.
Other aspects of IPv6’s address system enable a more standardised approach to addressing than is possible with IPv4, such as the use of a /48 prefix for most end-users, from large companies to residential users, with the capability of running 65,336 LANs, each with an essentially unlimited number of computers.
The Handbook describes the IPv6 extensions to the Domain Name System and the IPv6 packet header, with its system of extension headers for various applications such as Mobile IP and authentication.
We provide a detailed description of Mobile IPv6 and some enhancements to it, including NEMO (mobility for entire networks), Fast Handovers and Hierarchical MIPv6 Management (HMIPv6). We also discuss the difficulties of integrating mobile IPv6 and multicasting with cellular mobile systems and how this would compare with alternative arrangements, such as MediaFlo and DVB-H.
The Handbook includes a detailed discussion of SHIM6 - the IPv6-specific multihoming protocol which operates at the level of each host computer, rather than for an entire network via a router. There are several challenges to its successful operation, and questions about its usefulness when an entire network is to be multihomed.
The problems IPv4 is experiencing with unconstrained growth in the global BGP DFZ (Default-Free Zone) routing table - the Internet’s routing and addressing crisis - are likely to be repeated with IPv6 if and when IPv6 becomes widely adopted. We discuss attempts to avoid this through the use of address assignment policy and the filtering of BGP advertisements. Now that three RIRs assign Provider Independent /48 prefixes to end-users, we review the difficulty in constraining DFZ routing table growth unless a new routing and addressing architecture is developed. We also critique calls for the RIR’s control over IPv6 address assignment to be opened up to include multiple for-profit registration companies.
We review the transition mechanisms for running dual-stack host operating systems with IPv6 connectivity via tunnels over the IPv4 network. An IP “stack” is multiple layers of software which enable the computer to communicate via each Internet Protocol. A computer with both IPv4 and IPv6 protocol stacks needs both IPv4 and IPv6 addresses. Since IPv4 computers cannot communicate directly with IPv6-only computers, dual-stack is the only feasible method by which end-users could adopt IPv6 while still being able to communicate with all the computers in today’s IPv4 Internet. There are several approaches to this, including 6to4 and Microsoft’s Teredo. Yet only when almost all computers have IPv6 could most end-users do without IPv4 connectivity. This ubiquitous adoption of IPv6 will not occur in the foreseeable future, since IPv6 is yet to provide sufficient benefits for it to be adopted by significant numbers of ordinary end-users.
Finally, the Handbook considers the countries and application areas where IPv6 is most likely to be adopted, including mobile and IMS (IP Multimedia Subsystem) applications and in China.
For more info see:
