Re: [mif-arch-dt] Strawman solution proposals for DHCPv6 and RA support for multiple provisioning domains

[Tim Chown <[e-mail]>]

On 22 Nov 2013, at 21:17, Ted Lemon <[e-mail]> wrote:

> On Nov 22, 2013, at 3:38 PM, Marc Blanchet <[e-mail]> wrote:
>> But if is a human-readable for end-users, as you seem to assume above, then we enter into the languages, scripts, i18n, normalization and
> 
> This is not that hard, and we shouldn't avoid it.   Of course names need to be human-readable. So encode them in UTF-8.   (I'm oversimplifying, because you still need a unique identifier, and that's likely _not_ the name.)

I think there’s many potential scenarios for PVD-IDs, and to what or whom they may or may not be exposed to or required by, be the 'consumer' users or applications.

We seem to commonly refer to a user selecting an SSID and a PVD-ID following implictly from that; so the PVD-ID is implictly acquired, and the user may never need to see it. The user device may however have multiple interfaces, e.g. WiFi and 3G. I believe that iOS7 runs MPTCP across both wired and wireless interfaces for Siri, which is an interesting example.

Another example is selecting and using a VPN; there is again an implicit PVD-ID associated with the connection when the user initiates the VPN (or the VPN auto connects). We said at IETF88 that certain applications should be configurable to run only if a certain VPN PVD is available.

In the case of a mutli-homed IPv6 site with multiple prefixes, from different providers, users on workstations with two IPv6 prefixes don’t currently make selections, rather they rely on address selection or some other policy to be applied.  In this PVD case, an application may again wish to only run if a certain PVD is available, e.g. one prefix is associated with a PVD for some IP TV service, the other prefix is used for general Internet.

The above behaviours don’t require users to select PVDs, except where they are already exposed today as SSIDs or VPNs, but applications seem more likely to be able to assert a certain PVD is in place?

Perhaps we need something in the draft to discuss the possible ‘consumers’ of PVDs?

Tim