In the Internet at large, deployment of IPv6 has not progressed as fast as was hoped or expected when the initial concept of the IMS first came about. Many factors have contributed and continue to contribute to this, but in general any change in the core routing infrastructure of the Internet is likely to take a much longer time than one adopted at the edges of the network.
Network Address Translators (NATs) constitute a change that falls into the latter category. NATs are widely acknowledged as detrimental to new services as well as existing ones (e.g., SIP-based VoIP), but in terms of investment they offer a very affordable and easily deployable option to extending the fast-depleting address space of IPv4, thus slowing down deployment of the long-term solution in the form of IPv6.
 Actually, there are many other use-cases for deploying NATs, which have to do with security (firewall functionality), naming (e.g., private naming of hosts in a home-grown network), as well as addressing, just to name a few.
In many ways, what is happening in the Internet at large is descriptive of what is happening in the mobile domain. While IPv6 is seen as technically superior to IPv4 for the IMS, mandating IPv6 support represents quite a big barrier to deployment of any system. Using IPv6 exclusively also requires that roaming partners and service providers migrate to IPv6. This is why many early deployments have chosen ...