Re: [Ice] ICE Generation clarification questions

[Thomas Stach <[e-mail]>]

Peter,

thanks for the reply.
More inline


On 2017-02-27 22:53, Peter Saint-Andre wrote:
> On 2/27/17 2:13 AM, Thomas Stach wrote:
>> Hi,
>>
>> I'm a bit confused about the definition of an ICE Generation and how it
>> is used in Section 14
> To be clear for folks on the list, this is in reference to the trickle
> draft.
>
>>     Generation:  The complete set of candidates sent within an ICE
>>        negotiation session.
>>
>>
>>
>> Section 14 and the definition of Half/Full Trickle then uses terms like
>> "first generation"
>> "complete generation",
> For half trickle purposes, I think it would be best to use the phrase
> "full generation" or even "complete set of candidates". That is: under
> half trickle, in the initial ICE description the initiator sends all the
> candidates it might possibly send.
OK.
Since I learned from below that the generation is the union, the term 
"set of candidates" seems to be the better choice when it comes to 
describing what the initiator/responder sends.
>
>> "the responder can respond with an incomplete generation of candidates",
>>
>> "full generation"
>>
>> This seems to imply that the generation is not necessarily the complete
>> set of candidates,
> The generation is as defined above: the complete set of candidates sent
> within an ICE negotiation session. It would be better here to use the
> phrase "incomplete set of candidates".
Agree
>
>> but could grow during aICE Negotiation Session until end-of-candidates
>> is signalled.
>> So the generation rather seems to be the extensible set of currently
>> known/exchanged candidates.
> No, the generation is everything sent before an ICE restart (if any). In
> trickle the set can grow over time, whereas in regular ICE it can't.
Understood
>
>> It is also not clear to me if the candidates sent by the ICE initiator
>> and the ICE responder
>> belong to  different generations or if the generation is the union of
>> both candidate sets.
> It is the union.
Thanks for clarifying.
Thomas
>
> Peter
>
>