Converting to IPv6

By now most of you know that there is a new version of Internet addressing that has been introduced known as IP version 6 (IPv6). The process to integrate the new protocol into the network has already begun and it’s now time for smaller ISPs like my clients to begin looking at how they are going to make the transition. I call it a transition because the planned process is for the old IPv4 and IPv6 to coexist side-by-side until the old protocol is eventually phased out of existence. Some experts predict that the last vestiges of IPv4 addressing will survive until 2030, but between now and then every part of the Internet will begin the transition and will begin using the new address scheme.

The IPv6 specification makes major changes to internet addressing. Not only has the IP address length been extended to 128 bits but also the IP header format and the way header information is processed have been modified. Thus, transitioning from IPv4 to IPv6 is not going to be straightforward and it is going to take some work to go from old to new.

I think it is time to start thinking about how you are going to make the transition to enable both kinds of routing. Any small ISP will want to do this in a controlled and leisurely manner and not wait until there is an urgent need for it on your network. There are already some new kinds of hardware and software systems that are going to prefer to use the new protocol, and so small ISPs ought to get ready to make the change before you get a frantic call from a large customer asking why this doesn’t work on your network.

The basic process to get ready to migrate to IPv6 is to make certain that your core routers and other host systems in your network are able to handle IPv6 routing. There are three different transition techniques that are being used around the country to make the transition.

Dual-stack Network. This approach allows hosts and routers to implement both IPv4 and IPv6 protocols. This will let your network support both IPv4 and IPv6 during the transition period. This is currently the most common technique being used to introduce IPv6 into legacy networks. The biggest downside of the approach is that you must create a mirror-image IPv4 address for every new IPv6 address, and the whole point of moving to IPv6 was due to the scarcity of IPv4 addresses.

Tunnelling. This technique essentially hands off all new IPv6 routing to somebody else in the cloud. To make this work your network would encapsulate IPv6 packets while they are crossing your existing IPv4 network and decapsulate the packets at the border to the external cloud. This is somewhat complex to establish but reports are that it can work well when configured correctly.

Use a Translation Mechanism. This method is necessary when an IPv6-only host has to communicate with an IPv4 host. At a minimum this requires translation of the IP header packets, but it can get a lot more complicated.

And, as one would suspect, you can mix and match these various techniques as needed. It’s obvious to me that this could become very complex and there appears to be a lot of chances to mess up your network routing if you don’t do it right. Because of this we think it makes sense to start planning early on how you are going to make the transition. You do not want to wait until one or more of your largest customers are breathing down your neck demanding a transition, so you should start early and make a plan. Contact us at CCG and we can help you make a plan for an orderly transition to IPv6.

Leave a Reply