When it comes to deploying IPv6, you have two choices:
- Deploy a fully enabled IPv6 network only
- Run it alongside IPv4
However, is it feasible to run an IPv6-only network? Is anyone doing that? Probably not.
There are still a huge number of devices and applications that will only support IPv4, some of which will potentially never migrate to IPv6.
How can we manage the co-existence of two completely independent protocols?
When IPv6 stared to become reality a few years ago, engineers were faced with the challenge of coming up with a way to support the co-existence of IPv4 and IPv6 within a network. This led to three different lines of thought: dual -stack, tunnelling and translation.
When the time came to start testing and deploying the different methodologies, folks quickly found out that dual stack was the best option.
Tunnelling and translation aren’t bad options; they just aren’t as easy to deploy as dual stack, which only requires a few lines of IOS to enable. All that’s required on a dual stack router is that the interface have both the IPv4 and IPv6 addresses identified and voila! The router is dual stacked.
Most providers and vendors are going the dual stack route. I recently attended a Cisco webinar that looked at the performance of running dual stack through Cisco ASAs, and they proved out very little performance impact, even on long-haul transmissions.
What are the biggest challenges with a dual stack protocol?
Even though dual stack is the way to go, it’s not without a few minor issues. Applications must choose which IP protocol to use, which presents an issue when it comes to DNS. IPv4 DNS returns an A record and IPv6 returns an AAAA record. If AAAA (quad A) records are advertised to a host, you’d better have good IPv6 connectivity! When it comes to security, hosts listen to and secure both protocols, so even after heavy stress testing there are little to no performance implications.
What about you? How are you managing your transition to IPv6? Feel free to share your comments below.