[MMUSIC] Comments on draft-reddy-mmusic-ice-happy-eyeballs-06

[Ari Keränen <[e-mail]>]

Hi,

I think the new version of the happy eyeballs method looks better.

However, I was wondering does the Cmax in the example algorithm need to 
be negotiated or should some default value (range?) be recommended?

Also, wouldn't the IPv6 brokenness be discovered when the IPv6 server 
reflexive candidates are gathered?


And a few more specific comments on the text in the draft:

    The resulting checklist will depend on the priorities of the remote
    candidates.  It is not possible to ensure an even spread of IPv4 and
    IPv6 addresses unless both the remote and local sides uses the simple
    recommendations in this draft.

I guess it doesn't matter what algorithm is used, as long as it's the 
same and with same parameters.


    It is worth noting that there is a
    good chance it will some effect even if the remote side does not
    support this.

I'm not really sure what is the effect?


    (1)  HOST  IPv6 (1) Priority: 2129289471
    (2)  HOST  IPv6 (1) Priority: 2129289471
    (3)  HOST  IPv6 (1) Priority: 2129289471
    (4)  HOST  IPv6 (2) Priority: 2129289470
    (5)  HOST  IPv6 (2) Priority: 2129289470
    (6)  HOST  IPv6 (2) Priority: 2129289470

Is that 2x3 candidates with the same priority intentional?


Cheers,
Ari