Do we really need IPv6?

June 20, 2012

For over 15 years we have been hearing that the world would soon run out of IP addresses, and so, version 6 of the Internet Protocol was planned to save us all from the dwindling address space. We managed to survive throughout these years without really needing IPv6, thanks to the cleverness of CIDR (Classless Inter-Domain Routing) and NAT (Network Address Translation). Many claim that NAT is really unpleasant, since it violates the layered architecture of the OSI model, but it proved really useful, and the industry learned to work around the difficulties it imposes for SIP calls, and peer-to-peer communications in general. So, the question is: do we really need IPv6 after all?

In short, it looks like we will need it soon (and hopefully no one is going to write a review in 5 or 10 years wondering about that same thing!). In February 2011, the Internet Assigned Numbers Authority (IANA) allocated the last two free IP ranges to the regional internet registry in Asia. This does not mean that the IPv4 addresses are finally finished, since the regional registries still have unused addresses in their assigned ranges, but it is a good indication that what we have will soon not be enough. Given this reality, it is certainly time for network equipment vendors and service providers to consider preparing their software to work with IPv6, if they have not already.

This preparation should not be extremely complicated, since the low level details are taken care of by the operating system. However, when talking about 99.999% availability, it is not wise to rely too much on the OS. Our experience doing that for vendors has shown a number of critical issues that can arise and will need adjusting to work as intended.

To name a few:

• SIP Devices failing to register due to socket-level problems.
• Call agent refusing SIP messages with certain character combinations.
• SNMP messages being lost.
• Call failures associated with media negotiation problems between different end points (e.g., IPv4-only end point talking to IPv6-only end point).
• Possible security holes associated with tunneling IPv4 in IPv6 or vice versa.
• Vast amount of regression tests needing adaptation to work with the new protocol.

We probably still have a few years until we really begin to “need” IPv6, but it is certainly unwise to sit and wait to see what happens! Do you think IPv6 is necessary and what has your organization done to prepare the way forward?

References:
• IPv4 to IPv6 short guide: http://www.sizlopedia.com/2011/02/04/ipv4-internet-addresses-finished-short-guide-on-the-transition-to-ipv6/
• World IPv6 Day: http://en.wikipedia.org/wiki/World_IPv6_Day
• CIDR: RFC 1519, Wikipedia (http://pt.wikipedia.org/wiki/CIDR)
• NAT: RFC 2663, Wikipedia (http://en.wikipedia.org/wiki/Network_address_translation)
• SIP: RFC 3261, Wikipedia (http://en.wikipedia.org/wiki/Session_Initiation_Protocol)

FacebookTwitterGoogle+LinkedInShare
Categories: Technology Trends