Skip to Content.
Sympa Menu

wg-multicast - Re: multicast testing etiquette

Subject: All things related to multicast

List archive

Re: multicast testing etiquette


Chronological Thread 
  • From: Marshall Eubanks <>
  • To: "Richard Mavrogeanes" <>
  • Cc: "Frank Fulchiero" <>, "wg-multicast" <>
  • Subject: Re: multicast testing etiquette
  • Date: Fri, 28 Sep 2007 12:40:57 -0400

Well, also it is very rare to have a multicast network failure on an ethernet LAN.

So

- if you want to test codecs, encodings, broadcaster settings, etc., use unicast first
- if you want to test multicast config settings, like sdp files or proxy relays, do it on the lan the
servers are on
- use the network to test multicast routing once the first two are well in hand.

Of course, we are morphing more into debugging and trouble shooting...

Regards
Marshall

On Sep 28, 2007, at 12:33 PM, Richard Mavrogeanes wrote:

Right.

If it works going ten inches from Computer A to Computer B via unicast,
then it will work going from Computer A to millions of computers that
are just like Computer B, provided the network (i.e. multicast) is
right.

So there is Computer A / Computer B compatibility (the codecs), and
there is Computer A / Computer B protocol compatibility (the network).
Since you must use a network to test fundamental Computer A / Computer B
codec compatibility, I think it best to use the most common network
denominator which is unicast.

/rich


-----Original Message-----
From: Frank Fulchiero
[mailto:]
Sent: Friday, September 28, 2007 12:24 PM
To: Richard Mavrogeanes
Cc: wg-multicast
Subject: Re: multicast testing etiquette


On Sep 28, 2007, at 11:59 AM, Richard Mavrogeanes wrote:

Just one more comment to go along with Marshall:

Hi Richard, I'll take as many as you can give!


There is no need to employ multicast to sort out codec / player
compatibility issues. In fact, multicast (or the lack thereof)
tends to
confuse the codec issues, at least in the minds of the target viewers.
Why doesn't this work? Packet loss? Flooding? Routing? Or an oddball
codec?

I think I understand. Are you saying that if a specific codec/
implementation works in a specific OS/client software in http and
rtsp it will work in multicast?
That would also assume the SDP files, possible relays, and multicast
network are correctly configured.
So if it works in http/rtsp, and not in multicast, it must be an SDP
or multicast network issue, right?



You may be better off doing http:// and rtsp:// server first,
discovering the codec compatibility. Then, after you think all is
well,
releasing the multicast to the world community. Generally, there is a
small number of folks willing to look at your stream for testing,
so the
problem multicast solves (global distribution) is not needed anyway.

Good points, appreciate the feedback.

Frank Fulchiero
Digital Media Specialist
Connecticut College




Archive powered by MHonArc 2.6.16.

Top of Page