[opendtv] Re: News: The Internet revolution is about to be televised
- From: Craig Birkmaier <craig@xxxxxxxxx>
- To: opendtv@xxxxxxxxxxxxx
- Date: Tue, 1 Mar 2005 11:08:27 -0500
At 4:59 PM -0500 2/28/05, Manfredi, Albert E wrote:
>Wrong again, Craig. It looks more like a cable net
>because in order to route packets between neighbors, you
>need to travel up the MPLS path toward the center of
>the network (aka head end), rather than just switch at
>a neighborhood node. Just like cable.
You still don't get it.
This particular issue started out as a discussion of P2P connections
in the neighborhood.
As John pointed out, cable and IPTV networks ARE different. Everyone
in the neighborhood is on the same loop with a cable system. There is
NO technical reason why P2P could not work within this loop. IF the
cable system chooses to route stuff all the way back to the head end,
that is a technical decision, not an infrastructure issue.
The problem with cable networks is that they allocate only a small
portion of their spectrum for out of band data (control) and uploads
of IP data. In order to make P2P work in the neighborhood they would
need to dedicate more bandwidth to move data onto the local loop.
>And by the way, PONs would drive you to that solution
>anyway, because they are deployed specifically to avoid
>having active electronics distributed throughout the
>network.
Cable ALREADY has active electronics distributed throughout the
network. They are not deploying PONs. It is the IPTV systems that are
using various PON approaches. PONs are designed to deal with the
routing of services, and they are provisioned with adequate bandwidth
to the subscriber to route all services. Where switching occurs in
these networks is subject to design decisions as is true for cable.
Placing routers at the fiber to VDSL transition can be viewed as
being similar to the transition from fiber to coax in HFC networks,
but what happens downstream of those routers is very different.
>Which means, neighbors are fed with a wide
>band pipe, similar to the coax from a cable's HFC
>plant, and must then filter based on frame headers.
>And once again, in order to get a packet from your
>hoiuse to your neighbor's, that packet will have to
>travel upo the PON to the switch or router, rather
>than simply go from the curb back to the neighbor's
>house.
This may be true. It is a design decision.
>And where it's feasible in practice, they will. The
>simple fact is, what real world IPTV nets end up being
>more similar to cable plants than to your notion of
>TV over the public Internet. Whether for reasons of
>security, QoS concerns, or the nature of the cable
>plant itself.
We disagree.
>
>> To be successful the IPTV crowd must provide the
>> infrastructure to bypass the multi-channel
>> gatekeepers.
>
>So for example, IPTV nets will simply become another
>multichannel gate keeper. This is the reality, not
>the fantasy. Context is everything. Better re-read
>Doc's post, Craig.
I did. You're the odd man out.
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:
