IPv6: time to act, we're about to hit a wall

Share

Like most people, when considering a new version of technology devices or operating system, I expect either to have an incremental number besides the device name or a brand new name, e.g , from Smartphone4 to Smartphone5, or IceCup to CakeCream etc…

So when I was thinking about the new version of the IP protocol, “IPv5” (version 5) would be the natural name of the new version… but it is definitely IPv6. So It caught my attention, wondering what’s revolutionary behind this new version not to follow the classical way of naming and realize an “incredible” jump from a version 4 to version 6!

what’s driving this change in such a major way?

The number of public addresses. This alone justifies the jump. IPv6 really a boundless number of IP addresses. Nevertheless, in my opinion, another strong argument justifies the shift to IPv6 : the missing backward compatibility of the previous version, IPv4.

At the time IPv6 was designed in the 90’s, it was originally expected that IPv4 and IPv6 would run in parallel (a dual stack). It was also believed that the transition to IPv6 would take place right away, making the transition period and IPv4 address depletion non-issues. But it didn’t happen that way, and the missing backward compatibility of IPv6 to IPv4 combined with the ubiquity of IPv4 makes it more difficult for companies to adopt.

we’re about to hit a wall

IPv4 depletion is nearly here, and there is no other alternative than IPv6. However, Internet Service Providers will adopt and implement different techniques to cope with the depletion of IPv4 addresses. This will ensure continuity for users accessing IPv4 Internet content and services while IPv6 content and services are developed.

So the near future of the Internet will continue to include IPv4. This will be shared IPv4, or “NATted” with such technologies as CGN (Carrier Grade NAT) and IPv6. The Internet will have a mix of connectivity types with potential degradation of end-to-end applications that cross over from one numbering system to the other.

advice for enterprises

It’s no longer sustainable to take a wait-and-see approach to IPv6 adoption. For large enterprises, and particularly for multinational companies, this is a major concern. Enterprises have a number of different entities connecting to their network, including customers, suppliers, partners & remote employees. As these organizations and people (and their ISPs) upgrade their own networks, enterprises must be able to accommodate their evolution to IPv6. So from an end-to-end perspective, it is preferable to make content and applications also available directly on IPv6.

I have met several companies, especially large companies, where IPv6 can become a very large project. To demystify IPv6 adoption, my first advice is to scope and plan the project before beginning. This step is imperative to adopt IPv6 correctly. It is essential for an enterprise to define a strategic plan to ensure its [entire] communications network and services will be able to integrate IPv6 seamlessly in a cost-effective manner, and aligned with the defined priorities.

Secondly, I recommend reviewing and making an inventory of the company systems visible from or accessing the Internet (e.g., public web sites, extranets, VPN access for remote users, Software as a Service).

Thirdly, you should review potential security risks, such as uncontrolled introduction of IPv6 into the network, and avoid IPv6 over IPv4 tunnels to external systems, which can leave the network open to intruders. You should also review your security policy and add IPv6 related risks and mitigations.

I foresee that that IPv4 will continue to be present for until 2020, give or take a couple years. This means that companies need to plan for ongoing and new network requirements to ensure that both protocols are supported and user experience is not compromised.

…Oh, and in case you were wondering, IPv5 did have a brief experimental existence as RFC 1819. ;-)

resources and statistics on IPv6

industry links

image © Torbz - Fotolia.com

Christian Petrus

I'm product manager in charge of the international Internet product line and IPv6 product. I have been with Orange Business Services for 14 years, in the international environment working with multinational organizations.  I'm really enthusiastic with the innovation, technological of course, but also, in general, with new, fresh and innovative ideas and concepts. One of my main interests, maybe also driven by my experience, is to create and transform innovations into real efficient and cost effective value for our customers.