Let the process be evolutionary

Updated: Nov 14 2005, 05:57am hrs
Migration from IPv4 to IPv6 may seem like a simple technical issue in terms of infrastructure. It, however, is like having to run meter gauge coaches at broad gauge rail infrastructure spanning across the continents and criss-crossing each other! IPv4 to IPv6 migration is fraught with numerous challenges. One would appreciate that it is not about physical infrastructure, rather, the issues related to education, skill-set and management are equally more vital.

Any communication between two persons or devices for that matter, necessitates that each of these are uniquely identified, especially so while using a public system or network. Even in the physical world, though some localities, roads and even towns may be synonymous, if one writes complete address like one sees in a usual form, every home, office, building or a shop would have a unique address in the world.

Similarly, everybody connected to the internet requires a unique address. During the early evolutionary years of the internet, provision was made for about 4 billion IPv4 (internet protocol version 4) addresses and it was presumed to be inexhaustible for the foreseeable future.

The way the demand for internet is growing and the way it is becoming possible to use non-computing devices and interaction may take place even between machines, the demand of IP addresses is on the rise. Looking at this possibility, IETF (Internet Engineering Task Force) had begun work towards enlarging the total number of IP addresses more than a decade back. Also, rather than looking at small increments (even a single digit addition expands the total number of telephone numbers tenfold), it was decided to expand it to a great extent in one go and hence, implying possibility of zillions of IP addresses for every human being on the Earth!

Though IPv6 does promise certain improvements in QoS and security, only large-scale deployment would help in identifying any tangible benefits to the end user. Ultimately, the users would choose from a range of matrix spanning different QoS levels and corresponding costs. On the other hand, since the IPv6 address becomes public, this necessitates that users are not only aware of the possible security risks and also update themselves with the latest threats and plausible tactics.

Most organisations take just one or two internet connections at individual premises while the number of individual computers could be in thousands at such locations. Similarly, most of the residential users do not use the internet day in and day out whether they are on dialup or broadband access. In such a scenario, even if the ISP has less number of unique IPv4 addresses, the effective number of IP addresses is expanded in the same way that numerous extensions are enabled through an EPABX. This is called NAT (Network Address Translation) and is meeting the user expectation.

Few Indian ISPs have already deployed IPv6, though they are ready to offer services to the willing subscribers and almost all new equipment is IPv6 upgradeable/compliant. Hence, there is no need to mandate IPv6 addresses.

Having an open test-bed for IPv6 is a welcome step in the right direction. All the same, it would be sensible to make it compulsory for all e-governence applications to make them IPv6 compliant. For the physical infrastructure, it is possible to prefix a ceiling price for migration and the costs are likely to come down. Also , application development and support for IPv6 migration can be a very tricky, time-consuming and expensive challenge! This demand-side pull itself would create sufficient traction towards migration to IPv6.

The writer is secretary, ISP Association of India