At 05:32 PM 27/02/2007, David Conrad wrote:
Arano-san, On Feb 26, 2007, at 8:34 PM, Takashi Arano wrote:Site-level NATs have already contributed to address conservation very much, but not will do more.If you are an IPv6-only site and you want to talk to the vast majority of the Internet that is IPv4-only, how are you going to do it if you don't have NAT? I suspect the "dual stack" model chosen by the IETF and the fact that the IPv4 free pool will be exhausted before IPv6 is ubiquitous has guaranteed the pervasive deployment of NAT.
"predicting the future is easy - the hard part is getting it right!"I suspect that the relatively conservative decisions taken in the IETF over IPv6 technology has lead to a rather unusual outcome of what is, in technology terms, an incremental refinement in the technology base, but with an associated implication of disruptive deployment (i.e. lack of capability for piecemeal deployment and backward compatibility with the installed base). Normally, and the Internet itself is a good example, disruptive deployments are driven by disruptive technologies - i.e. the technology offers deployers significant opportunities that are inaccessible to the incumbent operators, and the emerging market is driven by these new opportunities and the investments in the new technology are made in the expectation of market displacement from the installed base.
NATs on the other hand are a more classic example of incrementalism - piecemeal deployment, externalization of costs, no change in existing investment values etc, where the relatively massive deployment of this technology in recent times should come as no surprise.
The predicted exhaustion of the unallocated pool of IPv4 addresses creates some potential for disruption in this industry. The actions of various industry actors would naturally tend to a preference for options that externalize associated costs as much as possible and minimize the potential impact to the value of current investments and service infrastructure. That would appear to make the outcome that David Conrad paints above, that of pervasive deployment of NATS as one that would appear to be a preferred outcome for many actors. Of course this would be a short term preference under the current set of constraints, but the longer term questions about the ultimate viability of such an approach, and the consideration of other factors make the longer term picture extremely speculative.
http://www.potaroo.net/ispcol/2007-02/address-paper.html contains some of my own further thoughts on this subject, for those who are interested in this topic.