Q. about master/slave determination in H.245

Kumar, Vineet vineet.kumar at INTEL.COM
Fri Aug 21 12:36:49 EDT 1998


My comments below.

vineet
> -----Original Message-----
> From: WanJiun Liao [SMTP:wjliao at EEE.HKU.HK]
> Sent: Thursday, August 20, 1998 9:13 PM
> To:   ITU-SG16 at MAILBAG.INTEL.COM
> Subject:      Q. about master/slave determination in H.245
>
> Hi..
>      I have some questions about master/slave determination in H.245.
> Perhpas these questions are stupid... but could you please kindly give me
> some help?
>
>      With H.245, we need to do master/slave determination to resolve
> conflicts between two endpoints which can be the MC in a conference.
>
> (1) If there is no MC is active and all entities are of different types,
> the priority to win the master/slave determination is MCU > Gatkeeper >
> Gateway > terminal. THis means. e.g., with Gatekeeper and terminal,
> Gatekeeper always wins the determination? then why bothering to
> set "statusDeterminationNumber" to be a random no. in that certain range?
        [Kumar, Vineet]  At the time of master/slave determination, the
entities do not know the "terminalType" of each other.

> or only when those entities are all in the same type, then such setting
> just makes sense? i.e., "terminalType" overrules
> "statusDeterminationNumber" ??
>
> (2) An MC that is already acting as an MC shall always remian the active
> MC. However, assume that there are only terminals participating in a
> conference, what if a terminal that won the determination already
> decides to leave the conference at some time during the call? Should we
> need to perform such master/slave determination again?, or the terminal
> can assign the other to be an active MC?, or  the terminal by no means can
> not leave the conference in the middle of the call?
        [Kumar, Vineet]  It is assumed that the MC will not leave the
conference. It is does, then the conference ends.

> Please advise... thanks a lot..
>
> Best regards,
>
> Wanjiun Liao



More information about the sg16-avd mailing list