[opendtv] Re: News: TV Braces for the Apple Tablet

  • From: Craig Birkmaier <craig@xxxxxxxxx>
  • To: opendtv@xxxxxxxxxxxxx
  • Date: Mon, 1 Feb 2010 09:05:13 -0500

At 6:54 PM -0600 1/29/10, Manfredi, Albert E wrote:

Craig Birkmaier wrote:


 Sorry Bert, but with IP Multicast, telco networks are far more
 spectrally efficient than ATSC MHP for video delivery.
By the way, even this is wrong, Craig, assuming you are talking about cellco wireless nets here. 

I am talking about wireless broadband - cellular is so passe' ...

in a few years all voice traffic will be VOIP.
But yes, I am assuming that future wireless cellular (as in cellular networks) will be able to broadcast popular content.
If you want to transmit IP multicast to individual cell phones, and NOT just use IP multicast within the wired backhaul nets between cell towers, and then broadcast the signal, you need to establish two-way links to each cell phone. Because that's how IP multicast works, Craig. What makes IP multicast "efficient" is that it reduces the traffic between IP routers. IP routers are in the backhaul network.
Now, for IP multicast to the cell phone, each individual cell phone needs to communicate using unicast transmissions with its local cell tower. Each individual cell phone needs to tell the cell tower which multicast group it wants to participate in, and it needs to keep doing this every time it gets a query from the cell tower. These are frequent, like once a minute at the least. 

And this is inefficient?
We're talking about a tiny fraction of the bits that are being broadcast, and it is not clear that once the link is established there would need to be any further communication unless the device moved to a new cell. Perhaps you (or I) are being a bit too literal about the term "IP multicast."
Obviously you do not need to have a two way link to receive an ATSC MHP broadcast. It is equally feasible for cellular networks to broadcast popular data streams as well. If the broadcast is encrypted you would need little more that a brief two way exchange to verify a key, and maybe not even that.
The MAJOR difference is that MHP from big sticks requires protection in adjacent markets - i.e. white spaces. This is horribly inefficient compared to the spectral reuse of cellular networks. And then there is the significant overhead to make MHP even work, not that cellular networks are without their own overhead.
Not only that, but the two-way link between each cell phone and the cell tower is unique to that cell phone. Its power is adjusted according to the distance from cell tower to cell phone, and it goes through a handover procedure as the cell phone reaches the edge of a given tower's coverage area. 

Yup. I noted that hand-offs between cells could add overhead.
So as you can see, the spectrum required in the wireless parts of the cell network is not used any more efficiently here than it would be unicasting that video.
Not true. You do NOT need to duplicate the content payload for every device tuned to a multicast stream, just the device handshaking. It would be quite feasible for a wireless broadband provider to use some of their spectrum for broadcasts and other spectrum for two-way data, including the necessary data links to tune into these broadcasts. 

Is this not how Flo TV works?
The advantage of IP multicast is only in the backhaul network. On the other hand, if the cell phone uses a broadcast-only chunk of spectrum, one which does not eat into cell tower's two-way link capacity, that's when you'll see the advantage. 

Exactly!
So it comes down to the same old argument about spectral reuse. Big stick versus small cells that reuse spectrum efficiently. 

Regards
Craig



----------------------------------------------------------------------
You can UNSUBSCRIBE from the OpenDTV list in two ways:
- Using the UNSUBSCRIBE command in your user configuration settings at FreeLists.org 
- By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word 
unsubscribe in the subject line.

Other related posts:

  1. Home
  2. opendtv
  3. Archive
  4. 02-2010