Re: [rtcweb] RTCWEB Data Channel Protocol: Label vs Protocol glare

[Christer Holmberg <[e-mail]>]

Hi,

>> The draft says:
>>
>> "Note: There is no attempt to resolve label glare; if both sides open
>>
>>                  a Channel labeled "x" at the same time, there will be 
>> two Channels
>>
>>                  labeled "x" - one on an even Stream pair, one on an 
>> odd pair."
>>
>> I think the text shall be extended to also cover the Protocol, i.e. 
>> both ends can try to open a channel for protocol X at the same time.
>
> And that should be fine.
>
> Having two channels with the same protocol is analogous to having two audio rtp m-lines.

...and sometimes that is what you want. But, in case you only want one channel for protocol X, the basic mechanism doesn't prevent a glare.

(Then, as has been discussed, one can do other things from preventing multiple channels for the same protocol to be created.)

>> Also, even if there is a label glare, it doesn't have to be for the 
>> same protocol. As far as I know, endpoints can choose whatever label 
>> value they want, and they don't even need to be for the same protcol 
>> (unless specific protocols define what label value(s) must be used for 
>> channels associated with that protocol).
>
> IIUC the point is for applications to manage the label namespace as they wish.

Assume that you are using application X, and I am using application Y. Both X and Y support protocol Z.

Now, unless protocol Z specifies what label value to use, X and Y may use different values. I don't know whether it really matters, but my point was that same label value doesn't by default mean same protocol :)

Regards,

Christer