On 25/07/2013 11:51, Randy Bush wrote:
the issue is that ipv4 is essentially gone
therein lies the problem: ipv4 is not essentially gone, nor do I foresee that it will be gone in my lifetime. Nor do I see that the RIR pools will ever stop garbage collection of ipv4 address space, which means that for the foreseeable future, they will continue to play a part in ipv4 address allocation. Also they can still serve a useful purpose as the registries of allocated space. With regard to 2013-03, we know that: 1. the RIRs will no longer be the dominant source for injecting liquidity into the available addressing pool 2. demand for ipv4 address space over time will continue to increase 3. there will be a market for ipv4 address space. The ipv4 address market will work if the market stays liquid, but 2013-03 creates an environment for full deregulation of the addressing market with almost no control mechanisms for handling problems [There is a 24-month reallocation freeze, but I don't think this is going to work because it will not affect the reality of real world allocation transfers. Consequently, people will flout it, and then we will need to remove it from the policy mechanism because it's causing the registry function to break down.] Liquidity will be generated almost entirely by existing address space holders releasing existing address space into the market. So the interesting question is: just how much slack space is there out there? Answering this will give us a good idea how much potential future liquidity there is, and consequently whether full deregulation is likely or not to run into liquidity constraints, given historical addressing demands over the last several years. This will give us a better perspective to decide on whether 2013-03 is actually a good idea. Nick