All things related to multicast

[Frank Fulchiero <>]

On Jun 20, 2008, at 11:45 AM, Zenon Mousmoulas wrote:

> OK so let me see if I understand correctly:
>
> The cause of the problem was not that a video stream (corresponding  
> to the SAP announcement) was somehow "relayed" to the SAP group  
> itself, as we first thought (at least I did). On the contrary, the  
> problem was caused by miniSAPserver sending otherwise "normal" SAP  
> announcements at the amazing frequency of 33KHz, instead of the  
> intended 0.03 Hz. Right?
>
> If that is so, and that is what happened in each and every recent  
> case of a "SAP storm", then how do we explain the different bit/ 
> packet rate, assuming everyone tried to set the suggested  
> sap_delay=30?
>
> However Frank also notes this behavior was not consistent. So, do  
> you have any ideas what else could have triggered it, apart from  
> this setting? A specific video stream type perhaps?


What might have triggered it for us is when we tried to send a SAP  
for an MPEG4 h.264 RTSP stream, and changed cfg to

[global]
sap_delay=30
interface=en0

[program]
name=Connecticut College Ch3
user=Connecticut College
machine=136.244.96.104
site=http://www.conncoll.edu
address=233.88.214.139
port=5432
address=233.88.214.139
port=5434
playlist_group=ConnColl miniSAPserver

This seemed to create the following stream in the SAP server, viewed  
in terminal

v=0
o=Connecticut College 16916 1 IN IP4 136.244.96.104
s=Connecticut College Ch3
u=http://www.conncoll.edu
c=IN IP4 233.88.214.139/255
t=0 0
a=tool:miniSAPserver 0.3.4
a=x-plgroup:ConnColl miniSAPserver
a=type:broadcast
a=charset:UTF-8
m=video 54325434 udp 33

I documented this to the list at the time, but there seemed to be  
less interest in SAPS then.

Frank