I was never on the front lines of any operator's customer service shop, so I suppose there might occasionally be complaints.  Of course, I'm sure that there are customers who complain about the telco stealing their brainwaves through the fiber...

Realistically, though, most customers have gotten used to the idea of line powered equipment on the premises to provide broadband and video services... DSL modems, cable modems, STBs, DTAs.  Power consumption of an ONT is probably about in line with these.

The bigger problems with local powering are standby capacity vs cost for the backup batteries and ensuring that the customer takes care of end-of-life replacement. 

Maybe some operator might start to factor customer perception about local powering into a business decision...  but then think about eliminating: cost of buying power for all those remote supplies, maintenance, training, liaison with the electric utilities, safety issues, and environmental compliance for dealing with all those lead-acid batteries.

BTW, Seven, IIRC there were three PON vendors in Verizon.  I will neither confirm nor deny that I ever worked for any of them ;-)

That is the gain.

The loss is bandwidth for the future.  There are other advantages and disadvantages of PONs and fiber remotes.

Ever heard of anyone disconnecting FiOS over Power issues?  Between me and Duh!, I suspect our equipment served almost all the FiOS users.  Never heard of that one.  Did have a lady knock an ONT off the side of the house with a sledgehammer over color.

seven

simply put, you're not gaining the wrath of the customers when their power comsumption goes up from an ONU  powering all the way back to the head-end in the case of an all fiber approach such as Verizon Fios.

comtech,

I think the question is what are you gaining via the centralization of that last few hundred feet of coax in cost in contrast with what you might gain if that last few hundred feet is fiber.

The biggest tradeoff to me versus any form of overbuild is that the entire area has to be converted at once.  This means converting hundreds of dwellings in a single day. That seems like a hard thing.

seven

This is a over cheaper approach to a packet data cable overbuild than running fiber to the home.All that is required is to eliminate the current passives and actives after the node,and as I may rehash without fair of beating a dead house,use the existing hardline cable to power the these new ONU-like taps at the pole,or peds.There is absolutely no need to have fiber-optic drops cables to homes as coax is quite capable of delivering data over 200 feet over more much the same as UTP over 325 feet!If you have a better solution, you need to state it,my friend,and with imperical logics to back you up.

So... you wreck out all the passives and actives in the plant.  Plus probably also have to replace splitters in the prem, along with CMs and STBs and any line amplifiers.  Overlash fiber (presumably point-to-point?) on existing hardline, install new actives.  So all that's left is the coax drops plus existing hardline and power supplies as power feed.  And the drops are still constrained to a few GHz of bandwidth.  And transitioning customers over is going to be disruptive.  And you still have those power supplies and actives in the field.

I'd be mildly skeptical about the business case for that versus overbuilding DPOE. 

Hear this. The current HFC  plant basically has one node that feeds no more than five line extenders, or LEs'.What the cable companies can do is to reck out those the current actives and passive components,and extend the fiber from the node to specially designed taps that can convert light into RF to the coax drop system.Now in order to power the taps, the existing hard-line feeder cables could be utiliized.This is in essence,FTTC.There is an alternative to this approach called RF over glass, or RFoG, which is really just semantics because it is no different from what Verizon's approach with Fios. I don't believe in subs having to foot the utility bill for power usage from these ONUs' installed in customer's home! On the other hand, coaxial drops would work just fine for an all packet data system because the main inhibiting factors are eliminated by havin a single node and no subsequent active components.

By restricting their plant to a node, optical splitters,and taps, the cablecos can realize huge savings in power comsumption, maintenance cost in terms of annual proof of performance testing and signal leak detection, and the confinement out outages to a single location.There will also be a significant less maintenance technicians to police the plant.Sorry, maintenance techs!

Without embarking on major plant upgrade as the one I suggested, packet data transmission over the existing infrastructure is doomed to fail.

The liabilities of HFC as a physical medium for supporting a packet data network  rather than a broadcast television distribution network are, in order of immediate significance:

1) the sub-split, which constrains upstream digital bandwidth to a small fraction of available downstream digital bandwidth, including notoriously noisy spectrum below 15 or 20 MHz.

2) the 6MHz channel plan, which causes many inefficiencies such as loss of statistical gain, extended channel change time, requirement for multiple receivers in some applications, inflexible channel lineups, and inhibiting schemes for avoidance of single frequency interference.

3) the reach vs attenuation curve of the coax at frequencies above 1GHz, leading to practical limits on usable bandwidth.

On the other hand, in terms of migrating to a physical medium optimized to support a packet data network,  an architect might look at the piece parts that might need to be changed out and see:

1) Components like node and amplifier plug-ins and head end equipment which are relatively easy to change out and relatively few in number.

2) Components like STBs, CMs, splitters and directional couplers which are relatively easy to change out, but are deployed in such large numbers in or near the customer prem as to require huge amounts of labor and logistics to change out.

3) Components like installed coax which are labor intensive to change out and installed in large quantities.

And therein lies the dilemma.  Solving the most urgent problem requires replacing every diplex filter and directional coupler (and perhaps splitters too) in the plant and prem at the level of the nearest FRU.  One could argue that the cost for doing this, and also breaking the 6MHz channel plan, would come so close to the cost of overbuilding with FTTP as to not be worth it.   One would have to do a pretty deep business case analysis to prove or disprove this hypothesis.

Note that there were several high split systems that were introduced and fielded, but never widely deployed, largely because of disruption to OSP.

I agree with you on the point of the HFC architecture still have some potential,but as you say, MSOs will have to make the decision very soon if they will continue to maintain it.To offer advance services and expecting a reasonable return on their investment is going to be difficult given the high maintenance cost of coax vs almost maintenance-free fiber.The challenge for most MSO to offer fiber beyond the node is replacing existing underground copper cable with fiber-optic in pre-existing plant.Aerial would not be much of a problem because all that is needed is to extend that part of the fiber from the node to the pole,and the running either fiber-optic or coax drop to the premises.Cable MSOs still have a card up their sleeve.......wireless spectrum that they have bought,and Clear WiMax.

Valid points from a traditional HFC view.  What Rouzbeh's interview highlighted was the need to make the bold moves and change our network artchitectures to allow a data network to replace the traditional RF network; inculding the headends or inside plant.  I recall working the mid 90's with USWEST in Omaha NE, or when Dave Pendegras (sp) got the crazy notiion to turn the low split HFC system into a 1 Gig Ethernet network which I believe is now WOW.  We have phased many upgrades from 350 Mhz, 450 Mhaz, 750 Mhz, 1 Ghz return, etc.. and the gains to be made have been tapped out.  HFC still has a very long life and imbedded value, its how we drive information on that physical plant that needs to be addressed.

Also, the technology advances made with Unverse and FiOS TV have leap frogged the traditional RF headends and MPEG set top approach and they place less burdon on the OSP; i.e. no EDFA lasers, no RF overlays.  They have the ability to now challenge Direct TV and Dish with their channel and OTT offerings which MSOs will have diffuclty keeping up with in an RF world.

Can't change where you are if you're not to looking elsewhere.