[opendtv] Re: News: The Internet revolution is about to be televised

At 6:57 PM -0500 2/26/05, Manfredi, Albert E wrote:
>That was a refreshingly good post on this subject, Doc.

Agreed! I agree with almost everything that John wrote. But Bert 
quickly steered the thread back into the ditch.

John wrote:

>  > So we end up with a VPN to each house, no real
>>  interconnect between neighbors until we get fairly
>>  high up the distribution tree in the network.  This
>>  negates the localization effect of a neighborhood
>>  PTP network and the advantage of that approach.
>
>It also looks much more like a cable network. And
>it's more compatible with a PON cable plant. I don't
>know whether the PON is a common solution for IPTV or
>not, however. Do you have an idea on this?

Wrong. It does NOT look like a cable network. At the core, cable 
networks are based on local loops in the neighborhood. EVERY home in 
the neighborhood receives the same bits; they tune into broadcast 
services, or they pull packets that are being routed to their home 
(that are ALSO routed to every home on that loop).

As John correctly states, the IPTV infrastructure routes only the 
packets being used by each home, to that home. There is NO comparison 
with cable here. The point at which the routing is exclusive to a 
single home is typically far upstream in the network, possible at the 
CO.

As cable migrates to more demand based services (as opposed to 
services that are broadcast to every home) the nature of the network 
changes, but the routing remains largely the same. The major change 
is the number of subscribers on the neighborhood loop. Today the 
average is between 250 and 500 homes per loop. As demand based 
services increase this number must be reduced. For a built out 
network where MOST of the traffic is demand based the statistics say 
that there will be about 25 homes per loop.

>
>>  Longer term, the amount of VOD content is just too
>>  large a catalog to have more than a fraction
>>  locally placed.
>
>Makes sense. Again, this makes the IPTV and cable
>plants look more similar.

Only in the back office. Ultimately, what John is suggesting is that 
cable systems will be getting out of the business of hosting VOD 
libraries. They will get INTO the business of hosting servers for 
content owners that will be connected to private networks with access 
to the full libraries.

Today the cable guys are the content aggregators. In the future the 
subscriber will be the content aggregator, and the local hard disk 
will be the key to making this work.

>
>>  Eventually, we are going to get to network DVR
>>  solutions to get the expensive HD out of the STB and
>>  move it into the network where we can both store
>>  more content and share the resource.  This is
>>  basically the same model as a PTP STB network, just
>  > moving the storage a little closer to the center.

>Here I disagree with John. The HD is not an expensive component 
>today, and it will just keep getting cheaper and bigger. Putting 
>this storage into the network does not solve anything. it just 
>changes the infrastructure requirements. You need more bandwidth 
>into the home since every TV will need to pull bits from the 
>metropolitan network rather than the in home servers. The most 
>efficient use of the network will be to cache programming form IP 
>multicasts, so that the content is available for local demand based 
>usage.


Local caching is the MOST cost effective compromise in terms of 
network infrastructure and capital costs for the system operator. The 
subscriber buys or rents the HD. IF the storage is upstream the 
system operator owns it and must charge for that. And then there is 
the issue of fair use. If the system operator gets into the middle 
then they must pay for the content and share any revenues that are 
generated. If they can charge ENOUGH this can be more profitable. But 
competition will make it virtually impossible to charge more for a 
service that caches in the network, versus one that caches in the 
home.

>Or, viewed another way, you move the in-home PVR just
>a little further up the network. Out of necessity, in
>part, because that last drop might not be wide enough.

You missed it Bert.

If the storage is upstream you need maximum bandwidth all the time. 
You may not use it all of the time, but you must provision for the 
worst case - whatever you define that to be. With local caching you 
may be filling the caches during off peak hours, and the bits may 
come in slower than real time, in real time, or faster than real 
time.  You may still have the same peak bit rate requirements if 
every set in the home is pulling bits in real-time, but the average 
utilization statistics will be VERY different.


>  > The content people are the bottleneck in this right
>>  now, so the home based HD will be the solution in
>>  the near term.  That will eventually change as they
>>  realize that their content is less likely ot get
>>  stolen if it is not sitting on a HD in the STB where
>>  it can be probed and prodded forever.

Yup. It is a question of control. The location of the media is 
irrelevant. IF people want to steal your content they will. But the 
winners will be the ones who figure out how to get people to play by 
the rules. It is amazing that Apple actually got the recording 
industry to TRUST millions of people to cache those bits locally on 
their computers and iPods...

Bottom line, people do not consume video content like music. Music is 
MUCH more vulnerable because people collect it an consume it over and 
over and over...

A system that provides transitory storage  along with the ability to 
transform those bits to a permanent storage medium is likely to be 
the winner. You can pay for a viewing window that might last a few 
days; if you want to keep it, you pay an additional fee to burn it to 
DVD.

>
>And that's another example of how the different
>networks will appear more similar. I would think the
>cable-provided PVRs would perhaps have to be rethought
>as well, for the same reason you mention. Or possibly,
>each PVR can encrypt for just the one monitor
>connected to it. Hmmm. But this would still be easier
>to hack, being in the home and all.

People are MORE interested in using the media than stealing it. Any 
system that attempts to lock the media to a single device is likely 
to fail in the marketplace. When you get an asset into local storage 
you should be able to watch it on any display in the home, and to 
move that asset to a portable device for consumption outside of the 
home. Reasonable DRM solutions will be accepted. Onerous systems will 
fail.


>So this is what I mean by "context." One cannot extol
>the boundless virtues of IPTV unless one has a clue
>of what the IPTV *and* the competing networks'
>universes look like in reality. Not just fantasy.
>That is what is sadly lacking in the trade press.
>

Apples and oranges. Cable has evolved the way it has because they 
have spent hundreds of billions building the plant and upgrading it. 
IPTV plants have the opportunity to take a different approach. This 
involves much more than just the physical infrastructure. It also 
involves the fundamental business model. Attempts to emulate cable 
and DBS are not likely to succeed. To be successful the IPTV crowd 
must provide the infrastructure to bypass the multi-channel 
gatekeepers.

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: