Randy Bush wrote: > this assertion is false, or disingenuous at best. > backbone service providers are turning it on at great > pain, much of that pain due to lack of support from > large router vendors. A few are working on lab efforts, and a very small number are offering service. Even if the transit ISP's had full up support today, that doesn't get the job done. The real issue is to convince the smaller ISP's that offer service directly to the end user. > you have been operating on that theory for many years and it has not worked. If one goes back to the origin of ngtrans, I was the one pushing hard to make sure that dual stack was the priority, and that all of that was specified before any solutions for corner cases. Even with dual stack as the primary approach, many network managers are expressing reluctance because they don't want to overlook the corner case that is important in their network. Independent of that, app developers see no value in building code for a network that they can't rely on working. It has only been recent work with 6to4 & teredo creating the perception of an end-to-end (in this case it only works at the end points) IPv6 network that has allowed apps to move beyond what they can do with IPv4. > > Bottom line, the end user doesn't know or care about IP or any version > > issues. They type names, or click on links, and magic happens. > > yep. as i said <http://www.ietf.org/> usually works pretty well. So why shouldn't port 25 work? > > It is our job to make that magic happen in more places, and as > > transparently as possible. > > no. it is the vendors' and operators' jobs. Ok, it is our job to create the consistency so that vendors can create products for operators to deploy and deliver services. In any case, just having a deployment in a few backbone routers doesn't get the job done. Tony