wg-multicast - RE: testing
Subject: All things related to multicast
List archive
- From: "Richard Mavrogeanes" <>
- To: "Marshall Eubanks" <>
- Cc: <>, <>
- Subject: RE: testing
- Date: Sun, 13 Aug 2006 19:51:40 -0400
Marshall,
I don't wish to be picky, I just want to make sure I understand. Isn't the
issue exactly zero, and not just "less" without RTP? I'm trying to construct
any case where it would matter.
Granted, RTP brings other advantages and we support RTP too (for unicasts).
/rich
-----Original Message-----
From: Marshall Eubanks
[mailto:]
Sent: Sun 8/13/2006 7:47 PM
To: Richard Mavrogeanes
Cc:
;
Subject: Re: testing
RTP expects the even/odd RTP / RTCP pairing. True, you can override
it, but
the standard RTP monitoring tools then won't work well. Of course,
this is less of an issue
if you don't use RTP.
Regards
Marshall
On Aug 13, 2006, at 7:00 PM, Richard Mavrogeanes wrote:
> The MPEG-2 is not RTP, it is MPEG-2TS.
> But since the port is associated only with the group, I don't see
> why it would matter what port number you select.
>
> /rich
>
>
> -----Original Message-----
> From: Marshall Eubanks
[mailto:]
> Sent: Thu 8/10/2006 1:16 PM
> To:
> Cc:
> Subject: Re: testing
>
>
>
> Dear Frank;
>
> Not seeing it from here :
>
> bash-2.05b$ rtpdump -F ascii 233.88.214.137/4444
> <silence>
>
> BTW, you should use even ports for RTP.
>
> The whole question of multicast ports is a vexing one. Of
course, you
> need to look at
>
> http://www.iana.org/assignments/port-numbers
>
> Most multicasters just pick some at random (and, remember,
most
> multicast is RTP,
> so if you pick port X, X should be even, and you will use X+1
for
> RTCP AND if
> you are sending audio + video, then you will be using X+2 and
X+3 as
> well.
>
> The trouble with just picking a random one is IANA will
eventually
> assign it (unless
> it is up high in the private address space).
>
> AmericaFree.TV currently uses ports from 7500 to 8090, but I
am
> intending to switch this to ports in the
> private port space :
>
> The Dynamic and/or Private Ports are those from 49152
through
> 65535
>
> I have proposed a number of times that the multicast
community pick a
> port range (or ranges) for
> streaming, either through IANA or just in the private address
space,
> but this has never seemed to find favor in Mboned.
>
> My current thinking is that multicast needs at least 1000
ports (I
> could see sourcing 1000 SSM channels
> from one server), and so why not pick, say, 60,000 to 61,000
? At any
> rate, that is my current thinking for
> AmericaFree.TV (which currently has 46 channels and thus
needs almost
> 200 ports now).
>
> (I will post this separately to the I2 WG mailing list.)
>
> Regards
> Marshall
>
> On Aug 10, 2006, at 12:53 PM,
wrote:
>
> > We are just getting started with multicasting, wondering if
anyone
> > can see our multicast VBrick streams
> >
> > MPEG2 at 233.88.214.138, port 4443
> > MPEG4 at 233.88.214.137, port 4444 for video and 4644 for
audio
> >
> > We are connecting to I2 through the Connecticut Education
Network
> > (CEN). Last spring we could see our SAPs and streams on
CEN, but
> > they were not making it to I2 for some reason.
> >
> > Also, we set up a special subnet just for I2 activities.
Just
> > wondering if most institutions keep the firewall completely
open to
> > all ports for this, or whether there is a set of "mutlicast-
> > activity" friendly ports suggested be kept open, to
specific IPs on
> > the subnet.
> >
> > Thanks,
> >
> > Frank Fulchiero
> > Digital Media Specialist
> > Connecticut College
> >
>
>
>
- Re: testing, (continued)
- Re: testing, Frank Fulchiero, 08/11/2006
- RE: testing, Richard Mavrogeanes, 08/10/2006
- Re: testing, fful, 08/11/2006
- Re: testing, Predrag Radulovic, 08/11/2006
- Re: testing, Frank Fulchiero, 08/11/2006
- Re: testing, Chuck Anderson, 08/11/2006
- Re: testing, Marshall Eubanks, 08/11/2006
- Re: testing, Dave Devereaux-Weber, 08/11/2006
- Re: testing, Predrag Radulovic, 08/11/2006
- RE: testing, Richard Mavrogeanes, 08/13/2006
- Re: testing, Marshall Eubanks, 08/13/2006
- RE: testing, Richard Mavrogeanes, 08/13/2006
- Re: testing, Marshall Eubanks, 08/13/2006
- RE: testing, Richard Mavrogeanes, 08/13/2006
Archive powered by MHonArc 2.6.16.