Why should the enterprise migrate to IPv6 and what impact will it have?
IPv6 is essential to the continued growth of the internet, due to its vastly expanded addressing space.
Early adoption sectors include markets with strong growth in residential broadband, such as India and China, cable TV providers looking to deliver IP entertainment content to the home and mobile operators growing their population of IP end points, such as smartphones.
Between them, these environments require hundreds of millions of new IP addresses that are simply not available in IPv4, forcing providers to use IPv6.
To successfully communicate with these IPv6-only end points, organisations should be planning to IPv6-enable their public internet presence in the 2012-2014 timescale.
However, organisations should hold back from migrating their internal networks, as most organisations use private IP addressing internally and are therefore not facing any address shortages and the risks and costs of migration to IPv6 are substantial.
Is the migration to IPv6 as straightforward as vendors claim?
Migration to IPv6 is not a simple process at all. It is a completely new protocol that touches not only the network, but operating systems, IT security models and even applications.
An IPv6 project therefore needs to be a cross-disciplinary IT initiative involving all parts of IT in a closely synchronised programme. In some cases, especially with older software, the IPv6 status may be unknown and testing may be necessary to see if the item has its own network protocol stack or is "well behaved" only using the protocol stack in the operating system.
Therefore the scale of the transformation effort is difficult to determine in advance. In these respects IPv6 migration resembles the year 2000 initiative, but without a global deadline. Operating systems and supporting services such as directories will probably need to be upgraded to the latest versions.
Even when products and services claim IPv6 compliance there will often be feature and/or performance differences between IPv4 support and IPv6 support that must be taken into account.
Is there a downside to migration?
There are several issues for organisations wishing to migrate to IPv6. Due to its larger header size IPv6 consumes more bandwidth to deliver the same useful payload than IPv4. This will depend on packet size but for a typical business can easily amount to a 10 per cent increase in LAN and, more importantly, WAN capacity.
For most applications, IPv6 will also consume more processor capacity to deliver the same workload. While this impact varies enormously from application to application, a 10 per cent increase in process capacity used is average. While this is unlikely to be noticeable on client PCs it can be very noticeable on servers.
Gartner estimates that a full transformation to IPv6 in organisations would consume about seven per cent of the total IT budget of the business during the year of transformation and one per cent of the total IT budget for several years thereafter.
Finally, IPv6 security is very immature, as there has been little IPv6 deployment and therefore few efforts to attack it. As with any new protocol, once it is widely deployed it will attract the attention of those wishing to compromise it.
In the labs, vulnerabilities have already been found in IPv6 implementations that were not present in the same vendors' IPv4 implementations.
Many of these challenges will reduce over time, as IPv6 security becomes "battle hardened", drivers are tuned to improve performance and we all deploy newer and therefore IPv6-capable versions of hardware and software.
However, we believe this all argues for an approach of minimum deployment, which in the short term means focusing on the essential task of IPv6 enabling of the organisation's web presence, but the active discouragement of more widespread IPv6 use within the organisation.
Neil Rickard, research vice president, Gartner