RE: stds-802-16: QoS Classes

[Leonid Shousterman <leonids@breezecom.co.il>]

[Notice: It is the policy of 802.16 to treat messages posted here as non-confidential.]

>Would the codec, echo cancellor, protocol stack, and jitter buffer add up
to
>about 50 ms for G.729 codec over IP?  I assume that MAC is responsible for
>the jitter buffer.

I guess it would be best if the jitter buffer is placed in the end-terminal.

If you insist that the Wireless Access device should be responsible
for eliminating jitter then it would be better to place the buffer at the
_wired_ interface because this interface is closer to the end terminal, 
thus it should not be a function of the wireless MAC.

Echo chancellor must be a part of the codec itself and the delay it
introduces
is the inherent delay of the codec itself. 

Best regards,
Leon.


> -----Original Message-----
> From:	Myers, William K. [SMTP:WMyers@spectrapoint.com]
> Sent:	Tue, July 27, 1999 9:31 PM
> To:	'IEEE 802.16'
> Subject:	stds-802-16: QoS Classes
> 
> 
> [Notice: It is the policy of 802.16 to treat messages posted here as
> non-confidential.]
> 
> Ref: 802.16sc-99/28, good writeup Arun.  I like the idea of aligning the
> QoS
> definitions.  Do you have a suggested partitionioning of delay for the
> wireless access link we are specifying?
> 
> Would the codec, echo cancellor, protocol stack, and jitter buffer add up
> to
> about 50 ms for G.729 codec over IP?  I assume that MAC is responsible for
> the jitter buffer.
> 
> By the way, the SYSREQ comment you posted to the net on 7/22 was not
> readable, like it was in binary in place of txt.  You may want to resend
> it
> in text form.
> 
> Regards,
> Bill