Sent from my iPad
2^128 is not 2^64 * 2. It is 2^64 * 2^64. It might also help to review the history of how we arrived at 128 bits. The original plan was to go to 64 bit addresses. The additional 64 bits were added solely because of the idea of auto configuration, so, in reality, it's not using up half the bits for host addressing so much as we doubled the address size to accommodate auto configuration. While I do not subscribe to the theory that IPv6 address space is infinite, I do believe that it is more than adequate to survive more than 50 years of liberal allocation and that there are very real benefits to liberal sparse allocations. Let's try liberal allocations as designed for a little while. If we burn through 20 /12s at the RIR level in less than 15 years, then I will be the first to admit we should consider les liberal a location policy. At that point, we will still have 492 /12s in the first 1/8th of the address space available.
Owen
|