On 18-mrt-05, at 4:04, Thierry Ernst wrote:
I second the fact that the deployment of any important architectural change should mandate a solution to both multihoming, mobility, security, multicast and so on. [here, I'm speaking about deployment, not the charter].
I'm sorry, but forcing a multihoming solution to address mobility and multicast as well is insane. There is no evidence that layer 3 mobility is even useful or necessary, and the repeated claim that so many devices will be mobile without addressing the question whether these devices will actually want or need layer 4 sessions to move from one address to another only supports my suspicion that mobility is an answer in search of a question. Interdomain multicast doesn't exist if we round down to a reasonable number of digits.
And regardless of the merits of these technologies, _I_ am not willing to work on multihoming if mobility and multicast must be supported in the first version of the protocol, because I can't spare the additional time and I'm unwilling to accept the additional risk of failure. I suspect the same is true for most others who've been active in multi6.
So please let's do this shim in peace and revisit the interaction with other stuff when there is something concrete to build on.
Yes, that was the clear consensus of the multi6 discussion and that's why the shim6 draft charter is the way it is. I see no value in restarting that debate.
It's very important that the charter excludes this discussion, so that when the WG starts it can write the shim6 specifications without distraction.
Brian (writing as ex-multi6 chair and ex-shim6 BOF chair)