|
Biting the IPv6 silver bullet
Even foresight has its limits. Unlike the shockingly
myopic "640K should be enough for anybody" statement (which Internet
folklore incorrectly attributes to one Mr William Gates), the task force working
on the Internet Protocol (IP) in the early 1970s provided for something like
4.3 billion IP addresses (via a 32-bit address space) in the current avatar
of IP on which the Internet runs—IPv4. Now surely that number should’ve
been enough for everybody. But the growth of the Internet has boggled even the
most prescient minds, and as IP-enabled devices extend to the realm of the mobile
phone and home appliances, 4.3 billion seems like a hugely embarrassing underestimation.
The problem is compounded by an inequitable
distribution of addresses, with the US hogging about 70 percent and the rest
of the world left to scrounge around for the crumbs. India has been allocated
about 2.2 million IP addresses, while China has something like 28 million.
About a decade ago, an announcement by
the Internet Engineering Task Force (IETF) stating that the world would run
out of IP addresses near about the year 2008, created a furore everywhere. The
solution suggested by the IETF was an evolution to IPv6 (also called IPng, for
‘Next Generation’). With a 128-bit address space, IPv6 would provide for 3.4
times 10 raised to the power of 38 addresses (3.4 x 1038); that’s 340 undecillion
(trillion trillion trillion) addresses—be assured, it’s a number way beyond
anything planet Earth will ever need.
But back then, the IPv6 protocols were
still being perfected. So network engineers came up with ingenious workarounds
to enable the re-use of addresses. One is DHCP, the Dynamic Host Configuration
Protocol, wherein a few addresses are dynamically assigned to a large number
of hosts of which only a tiny subset are online at any given point in time.
The other is NAT or Network Address Translation. Using NAT boxes, all the nodes
and devices within an organisation—up to 257 of them—appear to be a single device
to the rest of the network and utilise a single IP address.
With NAT in full flow, the spectre of address
shortages kind of faded, and so did the urgency to move to IPv6. But like any
honest-to-goodness workaround, NAT had its limitations and also created a new
set of problems. Without getting into technicalities, suffice it to say that
NAT and DHCP don’t really matter if all one wants to do is a bit of Web browsing
and e-mailing. But that was yesterday. Today, more and more businesses and individuals
demand a unique presence on the Net which is ‘always-on, equally-connected and
easily-reachable’ without being tied down to a single service provider. This
means unique IP addresses. Further, the 3G mobile world and IP-enabled devices
demand unique IP addresses to ensure true end-to-end, bi-directional communication—something
not possible when hundreds of devices are hiding behind a single IP address
using a NAT.
So, not surprisingly, the IPv6 chant has
been gradually rising to a crescendo, especially in countries like Japan, Korea
and China, which see real benefits in migrating to IPv6 because of the imminent
proliferation of 3G mobile telephony in their countries and the IP-enabling
of every device and the kitchen sink in their home. And in the process,
a not unpleasant fallout would be a reduction in US dominance of the IP address
space. Small wonder then that the Japanese and Chinese governments have fixed
deadlines for complete migration to IPv6 by 2005.
Of course with IPv6 it’s not just a matter
of addresses. In developing the evolutionary protocol, the IETF has ensured
that many of the inadequacies of IPv4 have been done away with. In the area
of security, there’s now innate support for IP Security (IPSEC), and a robust
authentication header for data packets that provides almost infallible authentication
and integrity checks, thus preventing host masquerading attacks (Beware, all
ye vile spammers!). Another major feature is address autoconfiguration, allowing
mobile devices quick acquisition and transition of addresses, and plug-and-play
connectivity for an unlimited number of devices in a network. With autoconfiguration,
mobility and the ability to identify each device individually, the application
development opportunities are mind-boggling. And best of all, there’s no need
for a stroke-of-midnight rip-and-replace—IPv6 migration can be gradual, and
co-existence with IPv4 can be indefinite.
And what of India’s IPv6 progress? We have
an IPv6 Forum India (ipv6forum.org.in),
which is active in holding an annual summit, generating awareness and conducting
trials. There’s also a moderately active discussion group with about 45 members,
but there’s a need to step up the momentum of IPv6 activity considerably, if
India is to be anywhere near the frontrunners in adoption.
IPv6 might not be the silver bullet panacea
for all internetworking ills. But it’s a bullet that we’d do well to bite sooner
rather than later. Any takers?
Val Souza, Editor
valsouza@expresscomputeronline.com
|
