x
Cable Tech

FTTP Bidders Slashing Prices?

Pricing has become the hot topic du jour following last week's public update on the fiber-to-the-premises (FTTP) request for proposal (RFP) from SBC Communications Inc. (NYSE: SBC), BellSouth Corp. (NYSE: BLS), and Verizon Communications Inc. (NYSE: VZ) (see Carriers Give FTTP Update). Though none of the vendors we contacted would speak about pricing on the record, rumors are afoot that some vendors priced systems below cost in order to remain in contention for a contract.

Privately, vendors say a well-worn practice called "forward pricing" has reared its head. Forward pricing involves pricing equipment cheaply today, based on anticipated volumes in the future.

"The ticket to the dance is that you have to be willing to drop your pants and savage your margins in the near term," says Kermit Ross, founder of Millennium Marketing and a veteran of vendor-carrier contract negotiations. [Ed. note: And past master of the mixed metaphor.]

What vendors are counting on at this point is that the carriers will eventually order their gear at volumes that will make the price cuts worthwhile. Secondly, the winning vendors might have first dibs on other network elements that need replacing, such as next-generation digital loop carriers.

But will carriers ever follow through and order as much gear as they promise at the bargaining table? Industry sources remain split on the issue.

Some say that the carriers' public statements have wed them to a huge FTTP rollout over time. Verizon vice chairman and president Larry Babbio has said that his company will reach 60 percent of its consumer revenue base with a fiber solution within five years.

But the other RBOCs involved in the big FTTP RFP haven't been so forthright about their deployment plans. "We can't make any deployment decisions until we've completed the RFP," says SBC spokesman Wes Warnock. "We're still heavily involved in the process."

That said, BellSouth and SBC will start first office application (FOA) trials during the second half of 2004, and only after that will they finally award an equipment contract. In greenfield applications, where the first FTTP rollouts will most likely occur, BellSouth expects to reach 135,000 new homes a year out of the 315,000 new homes in its territory, said Peter Hill, VP of technology planning and development at BellSouth, during a presentation at the United States Telecom Association (USTA) meeting last week.

Verizon is already said to have picked Advanced Fibre Communications Inc. (AFC) (Nasdaq: AFCI) for its first run of trials and deployments, according to analysts' reports (see Speculations Boost AFC). Not surprisingly, AFC's competitors quickly point out the likelihood that "aggressive pricing" factored heavily in the carrier's decision.

For AFC, having an installed base of digital loop carriers that can be upgraded to handle PON applications also helped keep its costs down. "Since carriers already deploy the [AFC] AccessMax, we know it works with current circuit switches (i.e., it supports standard interfaces TR08, GR57 and GR303, all of which are included in the FTTP RFP)," writes Merrill Lynch & Co. Inc. analyst Simon Leopold in an August 27 note to clients.

There is a historical precedent for forward pricing. In late 1996, Ameritech, BellSouth, Pacific Bell, and Southwestern Bell all signed multi-year contracts with Alcatel SA (NYSE: ALA; Paris: CGEP:PA) to deploy its ADSL DSLAMs and related equipment. Industry lore has it that Alcatel won the deal largely because of its ability to price its gear based on anticipated volumes. "The RFP was designed to help the individual companies meet growing customer need for high-speed data network access by creating volume orders for equipment, thereby driving down ADSL costs and speeding deployment," the companies said in a joint press release.

The effect of the huge RFP win was obvious. Alcatel quickly became the market leader in ADSL DSLAMs, delivering more than 400 DSLAMs in North America during 1998, according to RHK Inc. By the end of 1999, RHK notes that Alcatel held 51 percent of the ADSL DSLAM market, with no other vendor commanding more than 16 percent share.

While today's FTTP race isn't exactly analogous to Alcatel's early conquest of the DSL equipment market, the similarities are strong enough to make a case for forward pricing. Vendors say they were each asked to provide a five-year forecast at three different volume levels, but none will divulge the levels.

There are only two constants in carrier contract negotiations, vendors say. The first is that the initial price a vendor gives in response to an RFP is never the final price. The second is that, whatever the circumstances, carriers won't let a vendor bid at a higher price than its initial RFP response.

"If you win this, it could mean a lot of business," says one vendor bidding for the RFP. "If you don't, it means you could be shut out for a very long time."

— Phil Harvey, Senior Editor, Light Reading

photonsu 12/4/2012 | 11:17:18 PM
re: FTTP Bidders Slashing Prices? Bonnyman,
What I'm talking about is the future, which get closer each day. In that scenario, an active FTTC node that is fully equipped (loaded with VCSEL array optics) costs less than the per-port cost of the passive splitter. It gets installed once, and connected to the drop cables also installed at the same time to save deployment $$$. Drop cable is most likely pulled through underground conduit that carries your current copper drop, which seems to be common for homes built over the last 15 to 20 years. Provisioning then is a matter of activating the VCSEL link remotely.
In that future model, power to the FTTC node will be backfed to the node from subscribers over the unused or shared copper pairs. Just need a watt or two.
bonnyman 12/4/2012 | 11:17:19 PM
re: FTTP Bidders Slashing Prices? photonsu writes (regarding PONs):
"2) Bandwidth upgrade strategies tend to speak of reducing subscriber count on a splitter. This means one needs to:
(a) install new splitters (or pay for them when first deployed) at passive node to feed existing subscribers (down time and truck roll).
(b) deploy new fiber to feed additional passive node splitters. This may be 20Km in length.
(c) place an EDFA at the passive node?

3) With passive splitter nodes, there is no opportunity to do remote provisioning. So unless one is willing to give up on security, trucks roll to bring on new subscibers as take rate changes. So much for OPEX and cheap labor (which we all are these days)."


A truck roll will be required to activate the first customer at a new location regardless of the FTTX technology used. Someone will have to mount the NID on the side of the house and run a drop cable to it.

If, as previously proposed in this thread, there's no NID and the house is wired with fiber (and its' TVs, VCRs, PCs, etc. are fiber-ready), a truck will probably still roll to run a 100' to 200' drop cable to the house. (I think it's unlikely anyone will run the fiber drop cable to the house until the homeowner subscribes -- that's not the FTTH deployment model for overbuilds so far).

For PONs, quick trip by the truck to the splitter box shouldn't take more than a few minutes on the way. For active electronics in the field, it's likely a truck will still roll to the box every so often to stick in another interface card of some sort to handle additional customers. (I doubt the telco will fully populate the active nodes at first.)

If the house is sold and a new owner moves in, no truck roll should be required using either FTTH technology.

The biggest OPEX savings in FTTH is in maintenance -- fiber is more reliable than copper, especially when trying to run high speed signals over electrically deficient copper cables. Existing high pair count copper cables often have bad pairs and various defects that require troubleshooting, tweaking and pair reassignments before upgrading services.
photonsu 12/4/2012 | 11:17:23 PM
re: FTTP Bidders Slashing Prices? Bonnyman,
Thanks for taking the time to read my input and between the lines.
photonsu

I bash PON because I know it well, and it simply doesn't scale. At the risk of beating a dead horse:

1) Point-to-point lower speed optics will always be cheaper than that required for PONs. It isn't that PON transceivers are difficult to make or the special requirements are daunting, but that PON forces high cost where it can be tolerated the least. In/on our homes. I may be biased, but I do not see buying an HDTV with $M00 optic interface. P-P optics at $20 are already here.

2) Bandwidth upgrade strategies tend to speak of reducing subscriber count on a splitter. This means one needs to:
(a) install new splitters (or pay for them when first deployed) at passive node to feed existing subscribers (down time and truck roll).
(b) deploy new fiber to feed additional passive node splitters. This may be 20Km in length.
(c) place an EDFA at the passive node?

3) With passive splitter nodes, there is no opportunity to do remote provisioning. So unless one is willing to give up on security, trucks roll to bring on new subscibers as take rate changes. So much for OPEX and cheap labor (which we all are these days).

You said you don't get it! I think you do! It isn't that PON doesn't work. It's that on a large scale I don't think it will prove cost effective. This is why after 20 years the PON still struggles, and vendors dig deep with forward pricing. For me it is simple. What is best for all isn't the PON. We've all got to eat.

Again thanks!
bonnyman 12/4/2012 | 11:17:55 PM
re: FTTP Bidders Slashing Prices? photonsu wrote:

"Somehow I am not making myself clear. Let me try again.

What I am proposing is as follows:

One (bidirectional) or two (unidirectional) singlemode fibers from the CO to the neighborhood ONU as in traditional FTTC, fed with whatever bandwidth is appropriate for the size (subscriber cross-section) of the terminating ONU. CO link length can easily push 20Km and these optics are cheap and shared. An OC-48 will give 48 subscribers 100Mb/s each all-day and all-night.

Low power active ONU feeds subscribers using low power VCSEL array optics. Each subscriber gets a dedicated and private link that is capable of operating at GbE speeds (easy migration path) if needed as long as the drop length is <3Km, and 8Km if one is satisfied with 100/155Mb/s. (Currently I can only get about 33Kbaud, so 100Mb/s is like infinity and beyond). 850nm VCSEL and array technology leading to sub $20 per subscriber drop optics cost.
"

I wrote earlier against using multimode fiber, however, now that I better understand what you're proposing, I can see how this might work.

Having said that, the Bells are dead set against active nodes between their CO and the subscriber. This is because they don't want to power them. I don't necessarily see powered nodes as a big problem, but they do.

The irony is that the FTTC systems they are currently deploying require power. I don't know why they're inconsistent in this way. Certainly field-powered nodes have never bothered cable TV operators. Perhaps another reader can explain the reason for the Bells' aversion.

"So what about the NID? Totally passive! Just like the current copper NID. Nothing more than a passive transition point to the in-house drop cable. That cable will terminate on a fiber to whatever node you buy at Radio Shack (along with the terminated FO cable), Good Guys, Circuit City. WiFi, Tivo, PC, Fiber Fed HDTV, whatever. Simple fiber nodes would be less than $100.00. One will be able to buy what he wants just like you buy in-home equipment today. Once the Sonys of the world know FTTH is coming, there will be product on shelves to fill your every need, and at prices that all can afford."

This may be a good idea for greenfield developments in the future, but I think it's better to go with active NIDs for now. Rewiring a home is not a trivial matter. Take rates for FTTH services will go down if consumers have to rewire their homes when competing services (even if not as good) don't require it.

Even in a greenfield development, new home buyers will not want to replace all the existing home electronics they had in their old house.

"Notice no mention of triple-play. That is because we all know the world is going digital, so why if one is an RBOC, compete with old and costly technology that is destined for the junk pile or your basement if you are a techie."

Again, early deployments may require a triple play. People have lots of old TVs. They like having a cable TV lineup and surfing. It may take a while until video on demand becomes so pervasive, enough content arrangements are put in place and customers so comfortable with it that it's safe to totally junk the cable TV lineup.

Content in particular will be an issue. Example -- for at least another several years, NFL football will only be on ESPN. ESPN may require all of the ESPN channels be available 24 by 7 into the home. Or what about a polular show like Survivor that is supported by sponsors -- how would a video on demand system handle it?

Ultimately, I believe people will migrate to video on demand over IP. I would not want to bet against consumer behavior in the meantime.

You may be interested in a paper by Jim Farmer, CTO at Wave7:

"Video: By Broadcast, or IP?"
"Modern fiber-to-the-home (FTTH) systems are the first network architectures with the technical capability to deliver both broadcast video and IP video, if desired. This article outlines the two options and discusses pros and cons of each."

http://www.convergedigest.com/...

Farmer is one of the smartest guys in the cable TV business (and now the FTTH business).

Interestingly his technical approach to FTTH is somewhat similar to what you propose -- Wave7 is selling a system that looks like HFC (hybrid fiber coax) except with the "C" (coax) replaced with by fiber runs from an active powered node near the home.

"My conclusion, at the risk of being totally clueless is that PON technology is nothing more than a hold-over from the past, that only exists to protect the turf of the entrenched."

I see a lot of PON bashing on this board and I still don't get it. Municipal utilities, independent telcos and greenfield developers are all deploying PONs from companies like Optical Solutions. From everything I see, both the carriers and their customers are very happy with them.
sevenbrooks 12/4/2012 | 11:18:14 PM
re: FTTP Bidders Slashing Prices?
The major carriers in the US have broadly rejected FTTC. So in fact your fundamental assumption is flawed.

You are quite clear. Just the customers have rejected it. For 2 reasons:

1 - Cost of Power and the rest of the
2 - OPEX

You can put it any way you want, however you have clearly not dealt with OSP folks and their needs and problems. You have missed so many of the costs that its unreal.

seven
photonsu 12/4/2012 | 11:18:22 PM
re: FTTP Bidders Slashing Prices? Somehow I am not making myself clear. Let me try again.

What I am proposing is as follows:

One (bidirectional) or two (unidirectional) singlemode fibers from the CO to the neighborhood ONU as in traditional FTTC, fed with whatever bandwidth is appropriate for the size (subscriber cross-section) of the terminating ONU. CO link length can easily push 20Km and these optics are cheap and shared. An OC-48 will give 48 subscribers 100Mb/s each all-day and all-night.

Low power active ONU feeds subscribers using low power VCSEL array optics. Each subscriber gets a dedicated and private link that is capable of operating at GbE speeds (easy migration path) if needed as long as the drop length is <3Km, and 8Km if one is satisfied with 100/155Mb/s. (Currently I can only get about 33Kbaud, so 100Mb/s is like infinity and beyond). 850nm VCSEL and array technology leading to sub $20 per subscriber drop optics cost.

Now as I said, SMF is also a possibility. But it remains to be seen if 1310nm VCSELs and more expensive detector materials will be able to compete with the short wavelength component counterpart. I first bought Honeywell VCSELs at $6.00, Qty 10, back in 1998. So even if 1310nm does come on line will they ever be as cheap as 850nm parts currently selling in the 10s of millions per year? Given that they are likely to always be somewhat cheaper, is it enough of a cost advantage to make a difference? And do we need 1310nm to deliver the goods?

We need 1310nm technology mostly for reach if we accept my proposed architecture. And we get that whether one uses SMF or MMF. But if one accepts the hybrid FTTC followed by FTTH drops scenario, then do we need drops longer than 3Km or 8 Km from the neighborhood node?

So what about the NID? Totally passive! Just like the current copper NID. Nothing more than a passive transition point to the in-house drop cable. That cable will terminate on a fiber to whatever node you buy at Radio Shack (along with the terminated FO cable), Good Guys, Circuit City. WiFi, Tivo, PC, Fiber Fed HDTV, whatever. Simple fiber nodes would be less than $100.00. One will be able to buy what he wants just like you buy in-home equipment today. Once the Sonys of the world know FTTH is coming, there will be product on shelves to fill your every need, and at prices that all can afford.

Notice no mention of triple-play. That is because we all know the world is going digital, so why if one is an RBOC, compete with old and costly technology that is destined for the junk pile or your basement if you are a techie.

What price passive? Is the dream of PON reality? Every now and then one reads about ways to boost PON performance. Some have even suggested adding EDFAs to increase split ratios, speed or to amortize high PON costs over a greater subscriber cross-section. When I read these things I can only shake my head.

My conclusion, at the risk of being totally clueless is that PON technology is nothing more than a hold-over from the past, that only exists to protect the turf of the entrenched.

Just my honest opinion. Your turn.
dwdm2 12/4/2012 | 11:18:26 PM
re: FTTP Bidders Slashing Prices? Folks, much has been said and done about SMF vs MM fibers. In summary: as far as data transmission is concerned, Single Mode Fiber is the standard. "Why 155 mbps is not enough?" argument is not the way to justify using MMF, because, there is no end of this kind of questions -- Why not copper? Why not the way it is now? Why not the way it was 10 years ago? and so on.

Cheers
sevenbrooks 12/4/2012 | 11:18:27 PM
re: FTTP Bidders Slashing Prices?
Why is 155Mb/s too little?

If as we know that once leaving a serving area that there are 1,000s of subs per 155Mb/s what difference does it make to give them 1Gb/s for the last mile?

seven
bonnyman 12/4/2012 | 11:18:28 PM
re: FTTP Bidders Slashing Prices? bonnyman wrote:
"2. Those same dopants also give multimode its' higher attenuations (and thus reduced distances). If you read the entire Corning brochure, you'll see Corning is advocating it for premises links <300 m. (This is not likely to change in the future unless some new rare earths are discovered)"

to which photonsu replied:
"The 300 meter number is the range recommendation for 10Gb/s. If you go back and examine Corning press releases on the new MMFs, you will see that back in 1999ish they and Lucent with LazrSpeed demonstrated up to 3Km for GbE @ 850nm. My experiments demonstrated 8Km at 155Mb/s using a (1998 Honeywell) off the shelf $6.00 VCSEL."

See my earlier post citing Corning Cable Systems' specs for Infinicor multimode vs. their standard single mode fiber cables. Multimode fiber losses are 3.5 db/km at 850 nm and 1.0 db/km; single-mode losses are 0.4 db/km at 1310 nm. and 1550 db/km at 1550 nm.

10 Gig at 300 m -- much too short a distance for FTTH
1 Gig at 3 km -- still a short distance
155 Meg at 8 km -- too low a bandwidth, especially if it's shared with multiple users
bonnyman 12/4/2012 | 11:18:28 PM
re: FTTP Bidders Slashing Prices? bonnyman wrote (regarding single-mode fiber):
1. It's less expensive than multimode (and probably always will be). Multimode fiber needs careful (and possibly slower) preform manufacturing to get bandwidths as high as 2000 MHz-km. It also requires greater use of rare-earth dopants.
*****
to which photonsu replied:
Cost is a matter of volume. Whether MMF or SMF, the same amount of glass (and purity) is required as well as the time to deposit it. Your point on the dopant is valid, but it is my understanding that the incremental cost of the dopant is insignificant.
Todays differential between SMF and MMF is easily explained. The benefits in terms of I/O cost have allowed fiber manufactures to claim larger margins for MMF. To me it doesn't matter whether SMF of MMF wins, either way we achieve our collective goal.
Wasn't aware that Ge is a rare-earth.

You're right -- I was wrong. Germanium is not a rare earth.

Moving on to the rest of your post:

Perfect high bandwidth multi-mode fiber has a near-parabolic index of refraction curve. The fiber manufacturer lays down layers of soot on a rapidly rotating fiber "lathe" to build each layer of the preform. The thicker the layer, the more the index curve deviates from a perfect curve to something with a lot of steps trying to approximate a curve. This leads to lower bandwidths. To get higher bandwidths, you have take many more passes making thinner layers. (I've just described outside vapor deposition preform manufacture -- there are other methods, but all have the same issues of layer thickness and control to contend with.)

So far, making heavy machinery spit fire and rotate faster has not seemed to follow Moore's law.

Single-mode fiber has a much tinier core (<10 microns vs. 50 microns) and much simpler index of refraction profiles -- in some cases, just step changes between core and cladding.

As for purity, here is Corning Cable Systems' spec sheet for its' Altos outdoor fiber cables

http://www.corningcablesystems...:80/web/library/litindex.nsf/0/64bd202f3795f25b85256b6f007df95a/$FILE/CLT-93-EN.pdf

Attenuation for multimode fiber:
850 nm. 3.5 db/km
1300 nm. 1.0 db/km

Attenuation for single-mode fiber:
1310 nm. 0.4 db/km
1550 nm. 0.3 db/km

I'm traveling -- I don't have time to respond to the rest of your message for now.

A.B.
sevenbrooks 12/4/2012 | 11:18:30 PM
re: FTTP Bidders Slashing Prices?
In case you had a clue, its at the home. Its a fixed unit that is up or down, no provisioning, repair is replacement.

FTTC means boxes with plug ins. Every time you add a sub you go to the box. Nobody is preprovisioning ports for no service.

And the 20KM limit is the reach of the PON segment read the specs. Then take a look at the video overlay.

Sheesh please read before you post.

seven
photonsu 12/4/2012 | 11:18:32 PM
re: FTTP Bidders Slashing Prices? Seven,
In case you didn't notice, a 32 port PON has 32 sets of OSP electronics that are the responsibility of those concerned with OPEX. Why the 20Km limit? With a hybrid FTTC/FTTH approach one could go 80km with OTS long reach optics, and still drive 3 to 8 km fiber drops (depending on rate) to subscribers.
And in case you didn't notice, or were not paying attention, there is hardly any PON deployed out there, and what is, is only in areas not served by the by the big boys.
But let's stay to the point. CAPEX and OPEX needs subscribers. And if we don't do it right we will price ourselves out of the market and discourage customer service takeup. So what I am saying is we best not be telling lies and believing them, which I personally believe is the case with PON. Any PON.
VCSELs are here at both 850nm and 1310nm. One can use 850nm and 1310nm VCSELs over MMF where 1310nm operation will provide greater distance if needed beyond the above stated levels of performance. But you can't use 850nm over SMF unless willing to take 2dB to 3dB loss of a mode filter. But why fight each other. Our battle is with the copperheads. Put both SMF and MMF in the drop cable. The added cost is only for 100 meters or so, and is in the noise relative to the entire cable structure and installation cost.
Another myth that needs to be addressed is the idea that ones yard needs to be dug-up to install fiber. Over the last 15 or so years I've been paying attention to how services (gas, electric and phone) are being put in in new developments. In virtually all cases plastic conduit is installed to the curb so that those lines may be added as the contractor sees fit. My home is that way, with a 225 foot run to the pole. Could easily pull in a fiber pair.
st0 12/4/2012 | 11:18:33 PM
re: FTTP Bidders Slashing Prices? Let's do some guestimate:
(1) 50 micron MMF instead regular 62.5...(may be the Ericsson dual mode fiber = single mode center co-exist MMF).
(2) use surface emitting DFB instead of VCSEL
(3) No amp or with SOA
(4) keep it low speed of 2.5 instead of 10GBe
(6) CWDM 4-6 wavelength only
....
few km may be cheap and no problemo... all the technology are exist currently, someone just have to put things together (oh, boy, someone is going to worry about patent issues, "not invented here" is more of the block than anything else).

-st
photonsu 12/4/2012 | 11:18:33 PM
re: FTTP Bidders Slashing Prices? *****
1. It's less expensive than multimode (and probably always will be). Multimode fiber needs careful (and possibly slower) preform manufacturing to get bandwidths as high as 2000 MHz-km. It also requires greater use of rare-earth dopants.
*****
Cost is a matter of volume. Whether MMF or SMF, the same amount of glass (and purity) is required as well as the time to deposit it. Your point on the dopant is valid, but it is my understanding that the incremental cost of the dopant is insignificant.
Todays differential between SMF and MMF is easily explained. The benefits in terms of I/O cost have allowed fiber manufactures to claim larger margins for MMF. To me it doesn't matter whether SMF of MMF wins, either way we achieve our collective goal.
Wasn't aware that Ge is a rare-earth.

*****
2. Those same dopants also give multimode its' higher attenuations (and thus reduced distances). If you read the entire Corning brochure, you'll see Corning is advocating it for premises links <300 m. (This is not likely to change in the future unless some new rare earths are discovered)
*****
The 300 meter number is the range recommendation for 10Gb/s. If you go back and examine Corning press releases on the new MMFs, you will see that back in 1999ish they and Lucent with LazrSpeed demonstrated up to 3Km for GbE @ 850nm. My experiments demonstrated 8Km at 155Mb/s using a (1998 Honeywell) off the shelf $6.00 VCSEL.

*****
3. Single-mode fiber has extraordinarily high bandwidth. Most of the fiber being sold now has reduced or zero water peaks, allowing the use of twenty wavelengths each carrying 10 or 40 MBps data streams (using CWDM -- coarse wavelength division multiplexing).
*****
This is why I say get real.
Many are having problems trying to justify C/DWDM for use in the metro space. Even at todays low mux prices you are talking about adding at least $200 to the /subscriber cost with eithe WDM flavor. What is the alternative? SMFTTC, SMF(+)MMFTTH with <$20 VCSEL transceivers. As long as you are within 300M from the FTTC node, the migration path to 10G is there whether SMF or MMF is used in the drops.

*****
4. Existing cable TV, telco and power utility fiber already installed in the U.S. is 99.999% single-mode. The most likely locations to place a box with PON splitters probably already are served by singlemode fiber to an existing telco box or cable TV node. Using multimode downstream of the splitter would require running new fiber upstream and abandoning the existing fiber.
*****
See previous comment. No need to change anything.

*****
5. Less tangible, but probably a factor is that all the likely players -- telcos, cable TV operators, power utilities -- now use only single-mode and will likely resist any switch. This would be especially true of telcos.
*****
This "I will only use SMF" argiment is BS in my opinion, left over from the early 1980s when SMF came on the scene. Today the RBOCs have something like a half-dozen SMFs to choose from. Progress means change is always around the corner. I try to look at these things with an open mind. Hope you take my comments in that light?

sevenbrooks 12/4/2012 | 11:18:33 PM
re: FTTP Bidders Slashing Prices? Photon,

I guess you have not been paying attention. This is a FTTP bit and we are talking about 20 Km reach. Not 300m (at 300m I might as well do VDSL). Talking about such things is meaningless noise as they will not be developed and deployed. The whole point is to remove outside plant active electronics, NOT introduce more of them. Perhaps people are unclear on the dramatic OPEX that are intended through this.

seven
sigesux 12/4/2012 | 11:18:35 PM
re: FTTP Bidders Slashing Prices? Seven,
I've found that VCSEL's have some other problems related to thermal burst issues, this seems to especially true for the 1310nm devices, however they more than make up for these problems because of wavelength stability and lack the Multimoding of an FP laser, also the relaxation oscillation frequency is very high, all this makes the eye diagram degration look more random than for an FP laser.

VCSEL's also lend themselves to direct driving by CMOS drivers with 3V supplies, thats another $3 saving.

Anyway, with a 0dBm Tx VCSEL and a 32:1 splitter and 20Km feed you need about -33dBm receiver sensitivity. This is clearly very difficult especially for the OLT receiver. However when you look at the optical power budget over 15dB of the loss is in the splitter! If losses can be reduced by about 6dB than the solution is easy. Otherwise adding a FEC and viterby reciever is an option.

As for power savings in the ONT they are important because of battery backup requirements for (911 function) today you can buy a small Li-ion battery for under $1. The time that this small battery provides is proportional to the power used by the application. If a very low current POT's support mode was included than the system will meet all regulatory requirements. That saves about $10 on todays Power Management costs.

Note to all the bashers: I appreciate constructive feedback, but please to raise the level above "get real" and at least add some technical content to back your position.

optoslob
gea 12/4/2012 | 11:18:35 PM
re: FTTP Bidders Slashing Prices? "In terms of how long one has for clock recovery, its measured in terms of clock transitions not ATM cells."

Of course. The reason I state it that way is because I am assuming that the BMRx won't obtain a lock on the signal for a little while. Doesn't this assume that the first couple of ATM cells have to have no/dummy data in them? (Presuming that they'll be lost.)
bonnyman 12/4/2012 | 11:18:35 PM
re: FTTP Bidders Slashing Prices? Dear optoslob,
****
Low cost OLT APD receivers or possibly Raman pumping for increased OLT sensitivity.
****
Get real! FTTH will never happen if one doesn't learn to keep feet planted on the ground.
Check out the following link for a clue to the "real" solution WE all can afford.
Regards,
photonsu

http://www.corning.com/optical...


OK, let's do get real. The following is based on over a decade of working first with multimode fiber users, then telcos and cable TV and now power utilities:

The link cited is for Corning's higher bandwidth 50 micron Infinicor multimode fiber for LAN applications.

There are several reasons to stick with single-mode fiber:

1. It's less expensive than multimode (and probably always will be). Multimode fiber needs careful (and possibly slower) preform manufacturing to get bandwidths as high as 2000 MHz-km. It also requires greater use of rare-earth dopants.

2. Those same dopants also give multimode its' higher attenuations (and thus reduced distances). If you read the entire Corning brochure, you'll see Corning is advocating it for premises links <300 m. (This is not likely to change in the future unless some new rare earths are discovered)

3. Single-mode fiber has extraordinarily high bandwidth. Most of the fiber being sold now has reduced or zero water peaks, allowing the use of twenty wavelengths each carrying 10 or 40 MBps data streams (using CWDM -- coarse wavelength division multiplexing).

4. Existing cable TV, telco and power utility fiber already installed in the U.S. is 99.999% single-mode. The most likely locations to place a box with PON splitters probably already are served by singlemode fiber to an existing telco box or cable TV node. Using multimode downstream of the splitter would require running new fiber upstream and abandoning the existing fiber.

5. Less tangible, but probably a factor is that all the likely players -- telcos, cable TV operators, power utilities -- now use only single-mode and will likely resist any switch. This would be especially true of telcos.

Infinicor fiber is good stuff, but not for a FTTH application.

A major reason to use multimode in a premises fiber build is that many poorly capitalized "Mom and Pop" building wiring contractors don't have more expensive singlemode splicing equipment and they employ very low cost, less-skilled labor. That's not true in the outside plant world where telcos, cable TV operators, power utilities and their contractors own 1000s of single-mode fusion splicers.

Also, multimode fiber is easier to connectorize with lower cost connectors. Premises fiber builds are connector-rich with many patch panels, etc. Companies are frequently reorganized, offices reconfigured, LAN equipment moved, etc. Subdivisions -- except trailer parks -- are not often moved.

That's not true of the outside plant environment, which is likely to remain a splice-only environment for a long time. Some carriers are also uncomfortable deploying connectors in areas with wide temperature swings because of past issues with change in connector loss. Additionally, many lasers don't like the back reflections that come from some connectors.

A.B.
http://www.fiberplanners.com
photonsu 12/4/2012 | 11:18:36 PM
re: FTTP Bidders Slashing Prices? Dear optoslob,
****
Low cost OLT APD receivers or possibly Raman pumping for increased OLT sensitivity.
****
Get real! FTTH will never happen if one doesn't learn to keep feet planted on the ground.
Check out the following link for a clue to the "real" solution WE all can afford.
Regards,
photonsu

http://www.corning.com/optical...
sevenbrooks 12/4/2012 | 11:18:37 PM
re: FTTP Bidders Slashing Prices?
I have been looking at the VCSEL thing for about 3 years on this application. It all comes down to distance. Most people don't need the launch power for 20 - 40 KM for most applications. 5 - 10 KM reach for most customers should work. The issue is how to deal with the logistics of 2 different Laser Models of the same NID and how to know what to install. Other than that I see long wavelength VCSELs as a key potential cost reducer.

Interesting idea on the OLT receiver. Bit worried about the noise, but I will ponder it.

Not sure that the power savings at the ONT may be worth the cost. Cost is the primary concern. Lower power can save BOM cost, but not always.

I generally concur with your comments on the OLT burst receiver. However, I am not sure that cost is as important in this bit of the solution as is ruggedness. A more stable solution for CDR may be better than a cheaper one, unless the difference is dramatic. The cost of this circuit compared to the total BOM cost of the OLT blade is not a large percentage.

seven
sigesux 12/4/2012 | 11:18:37 PM
re: FTTP Bidders Slashing Prices? Knowing the ONT ID in advance does not do you much good because most system implementations try to keep the TIA simple so that it can be mounted as a ROSA. This means it is difficult to communicate the ID back to the TIA, more importantly the high and low levels are not constant during the burst. so the TIA only trys to get an approximate offset adjustment. The real signal recovery happens in the LA / CDR stages.

The APON spec allocates 10 cycles at the begining of the burst for LA level recovery and CDR locking. There are various schemes for fast clock recovery in the Gbps range look at 1394, GigE or USB2.0 systems.

As for Optical PON systems vs Magnetic HDD comment, both systems look like a variable current source to the electronics, the main difference is that in HDD systems the eye is most impacted by intersymbol interference whereas in the PON systen it is fiber dispersion and relaxation oscillations.

Anyway, I'm not here for a pissing competition. Bottom Line is, the socket goes to whoever figures out how to implement this function cheaply.

I'd rather use this forum to discuss specific PON problems. My list would read

ONT power saving (target 100mW sniffing mode)

PON Systems using sub 1mW ONT 1310nm lasers (VCSELs)

Low cost OLT APD receivers or possibly Raman pumping for increased OLT sensitivity

Regards
optoslob






sevenbrooks 12/4/2012 | 11:18:38 PM
re: FTTP Bidders Slashing Prices?
gea,

Far be it for me to defend him, but my guess is that he meant CD ROM drive or perhaps DVD drive.

In terms of how long one has for clock recovery, its measured in terms of clock transitions not ATM cells.

I think people are highly optimistic about the deployment of very high rate PON systems. If the bulk of video traffic is done on the overlay, I just do not see the bandwidth driver to upgrade. The impetus to GPON is multiple personalized HDTV streams to the home.

seven
gea 12/4/2012 | 11:18:41 PM
re: FTTP Bidders Slashing Prices? "I dont agree that upgrading APON to GPON will be that difficult because any Hard disk drive has a Gpbs burst Mode receiver."

Uh...huh? A hard disk drive is magnetic. You do not use a laser to read it. With a PON we're talking about optics and OPTICAL clock recovery, etc... this is not a trivial problem, and as of 3 years ago true burst mode receivers operating at 622 Meg we're just about state of the art.

Of course, as was pointed out, we don't need true burst mode for PON...the OLT knows in advance (via the ranging protocol) at least what optical power levels to look for. It probably needs to re-recover the clock I imagine (does this imply the first couple of ATM cells are blank?)

Your previous comment made a modicum of sense, however. I agree that A-PON is transitory and G-PON will be what actually gets deployed (if/when a PON has to be deployed, that is).
sigesux 12/4/2012 | 11:18:43 PM
re: FTTP Bidders Slashing Prices? There are two difficult receivers in a PON system, One is the ONT video reciever, this is very difficult because of the CTB/CSO requirements from a relatively weak signal. Also it needs to be cheap.

The OLT Rx is a typical Burst Mode receiver, usually the TIA has a Peak High / Peak Low detector that is reset by the PHY signal NEW_ONT. QB implements this different from Lucent and Fujitsu you can check out the pattents they are all several years old now.

I dont agree that upgrading APON to GPON will be that difficult because any Hard disk drive has a Gpbs burst Mode receiver. So there are plenty of semi companies that know how to do this chip.

Optoslob
sigesux 12/4/2012 | 11:18:43 PM
re: FTTP Bidders Slashing Prices? You guys seem to be forgetting that PON specs allow for the upgrading by CWDM overlay. If a particular client wants secure comms in both directions than allocating a seperate 155x nm laser at the CO and some passive filters at the splitter and the whole system is secure and still completely compatibe with APON.

In my opinion APON/BPON are only for trials, GPON is the system that will see real deployment.

optoslob
sevenbrooks 12/4/2012 | 11:18:45 PM
re: FTTP Bidders Slashing Prices?
By the way, I keep reading about how wonderful it would be if we did point to point fiber on this thread. That 30Mb/s per home down and 5 Mb/s per home up is to little.

There was an old Telco rule of thumb that 80% of traffic was local. The Internet created a rule of thumb that 80% of all traffic is remote. So, do you honestly believe there is anything like 4Mbps per sub leaving any given area? Lets pick a 10,000 person town for example 10,000 * 40 Mbps = 40 Gbps. If the rule of thumb is true (always need to look at the assumptions), then the bottleneck will not be the local loop. In fact, this 10,000 person town would require approximately 1 Gbps to get 128Kb/s DSL upstream out of town.

So what we are talking about is how oversubscribed the backhaul facilities are out of a town. Do you think they will be upgraded when even PONs are brought into the area?

seven
gea 12/4/2012 | 11:18:46 PM
re: FTTP Bidders Slashing Prices? "The ONTs receiver, while having to deal with a variation of absolute level has a constant source and a constant level. This looks very similar to a SONET reciever."

Damn. If I keep hammering these boards someone normally responds who actually seems to know what they're talking about!

So in other words, an ONT Rx is actually 'close' to a SONET Rx, except it doesn't need to re-learn what a zero is every time an OLT transmits, it's more or less remembers that OLT's zero from the ranging function.

Interesting. This changes my perspective slightly. We're still not talking about an off-the-shelf SONET Rx, though, but it might not be super-hard to upgrade the bitrate in future PONs.
sevenbrooks 12/4/2012 | 11:18:47 PM
re: FTTP Bidders Slashing Prices?
Gea,

The burst mode receiver at the OLT end of a PON connection is the hardest bit of the design from a Hardware Standpoint. 2 big issues:

1 - Recovering a clock rapidly
2 - Fast AGC on the signal

However the designs have a "ranging" function that is part of the addition of a new ONT to the PON. OLTs grant access to ONTs and therefore know who will be replying at any point in time. The ONTs receiver, while having to deal with a variation of absolute level has a constant source and a constant level. This looks very similar to a SONET reciever.

seven
integrator 12/4/2012 | 11:18:48 PM
re: FTTP Bidders Slashing Prices? Gea

The end user location transmitter for PON needs to power up and provide a synch signal and then the overhead bytes and then followed by the payload. For ATM this can be very small since the IP payload is in Ethernet bytes from the PC NIC card is too large to fit into a ATM fixed cell size. Since only 32 ONT's per PON the transmission window could be for 1/32 of a second for each, if there was no gap required between ONT lasers sending; but in partice this gap is never eliminated. The gaps can take 10-30% of the upstream PON transmission time depending on the quality of the transmitters and receivers used. Usually the ONT transmitters are cheaper products and the OLT is a higher cost to compensate.

With EPON systems the ONT Tx can be on for a longer period since the payload usually is not Segmented and Resegmented (SAR) to be broken into many small payloads like ATM. Thus the payload for EPON systems is usually supportable from 64 to 1318 bytes (packet size varies with each trans mission) to handle most Ethernet frame sizes to acheive very high throughput of customer payload with minimal delay. Jumbo Ethernet frames would not be supported without SARing if the EPON byte size limit is 1318 bytes.

If limted to this size of tranmission window per ONT the OLT Rx would need to follow the first transmission then have a gap with no transmission and be ready to synch up to a new ONT signal at a new power level with a short tranmission burst coming at it. These are Burst Mode Transmitters at the ONT so the OLT Rx must be built to handle the variation of power levels and the short bursty transmissions. These are special transceivers. The ATM and the EPON standards use 1490 nm lasers at the OLT and 1310nm lasers at the ONT.

There are several manufactures making these, Infineon is an example. Both ATM PON and EPON standards have very specific requirements for these transceivers.

Integrator
gea 12/4/2012 | 11:18:48 PM
re: FTTP Bidders Slashing Prices? "So there is zero technical issue with variations in signal level."

Well, that's not the way I understood things. Yes, normal SONET Rxs have an auto-zeroing function, but this takes place over a relatively long period of time (from what I understand it can be as long as a frame or two).

In this case each end user has a small fraction of a single SONET frame into which they can transmit, so that the "zero" is changing every several hundred bits or whatever. This means the Rx has probably no more than an ATM cell or two to figure out that the far-end Tx has changed. This means the PON Rx is very special, and very hard to upgrade (it's not exactly burst-mode, but it's close).

Then again, I don't claim to be a PON expert, so maybe there's some deeper logic or error correction that can be applied to overcome this difficulty (ie, of a nearby 0 being stronger than a far-away 1). But if someone's going to correct me, let them be the kind of person who has to spec and buy PON Rx's...they will know if they can merely buy off-the-shelf SONET transceivers or do they have to be special PON transceivers? (Or non-transceivers, since downstream and upstream are two different rates.)

bonnyman 12/4/2012 | 11:18:49 PM
re: FTTP Bidders Slashing Prices? In response to a very detailed message regarding ATM PONs from "integrator", palomar7 wrote:

"Bogus "analysis" integrator. Filled with flaws, inaccuracies, and twists of the truth. Not even worth correcting them."

Palomar, I was interested in what integrator wrote about ATM PONs and I am also interested in what you think the flaws are in his message -- could you elaborate with some specifics?

Thanks,

A.B.
whyiswhy 12/4/2012 | 11:18:49 PM
re: FTTP Bidders Slashing Prices? OK, I'll take that one, and say meetoo will not keep them in business. The best way to do that would be to fight in Congress or just move on to the new frontier: wireless. Separate the business into two parts, and let the old part whither and die. Use it as a tax writeoff.

Copper: Can it say Moo? If so, keep it. Otherwise, dump it.

-Why
whyiswhy 12/4/2012 | 11:18:50 PM
re: FTTP Bidders Slashing Prices? OK, looks like I stirred the pot.

gea, I respect all your posts except those to BM, but you (of all people) certainly know that it is common practice to make decision circuits self leveling...many FEC chips support "soft" decision making. So there is zero technical issue with variations in signal level.

While I am at it, these same chips make it possible to reach to within a hairs breadth of Shannon's limit for the chosen coding scheme. That's pretty impressive.

-Why
gea 12/4/2012 | 11:18:50 PM
re: FTTP Bidders Slashing Prices? "ATM PON doesn't use SONET framing. The downstream is broadcast to all ONTs and the upstream from all ONTs must signal one at a time. There is no way a SONET framer is going to support this mode of operation. In addtion the signals use Voice Circuit Emulation (AAL2/AAL5) not SONET TDM voice transmission."

OK, something's wrong here. Downstream SONET framing is no problem: If we assume an ATM-over-STS-12c, then the end terminals merely pull out the appropriate ATM cells. (I of course knew that the services were cell-mapped and not circuit-switched.)

Upstream I don't see why you believe there's no SONET framing. If there's any framing at all then it might as well be SONET (and I really doubt the OC-3 and OC-12 bitrates are mere coincidences).

Again, like a lot of people, you are confusing circuit-switching with SONET framing.


GPON, however, is not an ATM-PON from what I understand...from the last I heard it was going to leverage Ethernet and not ATM. That will not use SONET framing. ATM-PON, however, does use SONET framing.

teng100 12/4/2012 | 11:18:52 PM
re: FTTP Bidders Slashing Prices? TCP/IP is not particular interesting when we talk about the bandwidth usages, this only will limit the future applictions and the so called home-networking, which may have large bandwidth usages
than only one PC station.
integrator 12/4/2012 | 11:18:52 PM
re: FTTP Bidders Slashing Prices? GEA,

Here are a few corrections to the descriptions you gave for how PON operates.

Upstream PONs can implement Dynamic Bandwidth Allocation (DBA) for upstream traffic. This is a method where the ONT tells the OLT how much data it has ready to send upstream (buffered) and the OLT sends a message to the ONT to tell it when to start the transmission and for how long. This method allows for the ONT to send upstream for the full time period allocated by the OLT. When not all ONTs need to send information upstream at the samne time this allows most of the upstream transmission time to be used by the ONT that needs to send data. When GPON or EPON systems are used rather than BPON (155Mbps upstream) they can sent at 1Gig Speeds upstream. Thus up to 10 ONTs at 100Mbps can be supported at full 100 Mbps transmission throughput. Not much delay or buffering is needed for this to work.

The PON receiver in the OLT must be able to handle the variation in Power levels from many ONUs at diffent distances, it is designed to distinguish the signal variation from noise (but this is a shared resource and can be a more expensive transceiver). The optical transmission rate is 10 to the minus 12 signal quality over PON fiber networks. This is without Forward Error Correction software. PON is very capable of near error free transmission at this quality even with 20 Km distances with 32 ONTs.

PONs which start out at 32 ONTs can be split into two PONS of 16 ONTs. This is especially possible if the fiber placement is designed for this future transition step. With Gig speed PONs, most of the issues of bandwidth are eliminated. At 16 split Gig Speed PONs the issues of bandwidth on PONs disappear. 8 split PONs provide guaranteed bandwidth of 125M bps downstream and 100Mbps guaranteed upstream per ONT.

See the earlier posts I submitted for bandwidth and the difference between best effort and guaranteed speeds for ONT services.

Integrator
integrator 12/4/2012 | 11:18:53 PM
re: FTTP Bidders Slashing Prices? gea

ATM PON doesn't use SONET framing. The downstream is broadcast to all ONTs and the upstream from all ONTs must signal one at a time. There is no way a SONET framer is going to support this mode of operation. In addtion the signals use Voice Circuit Emulation (AAL2/AAL5) not SONET TDM voice transmission.

Take a look at the ITU GPON standards protocol stacks.

Integrator
integrator 12/4/2012 | 11:18:53 PM
re: FTTP Bidders Slashing Prices? rs50terra

Not true on splitter location for PON.
Since the lasers are at launch into the fiber near +1.0 with 1490nm laser their are not issues which require the splitter to be close to the laser. Splitters can be place up to 20Kms without any problems.

Integrator
integrator 12/4/2012 | 11:18:53 PM
re: FTTP Bidders Slashing Prices? Palomar7

Present you arguements.

The information I posted is correct.

Integrator
palomar7 12/4/2012 | 11:18:55 PM
re: FTTP Bidders Slashing Prices? Bogus "analysis" integrator. Filled with flaws, inaccuracies, and twists of the truth. Not even worth correcting them.
integrator 12/4/2012 | 11:18:56 PM
re: FTTP Bidders Slashing Prices? data_guy

Not true that PONs are limted to 30M bps. PONs are not as flawed as you put on.

PONs deliver more than 30Mbps to each home. They actually can deliver more than P-P fiber networks when the ONT has multiple data ports and or has a GE interface data port. Try thinking of a PON as a switch/router at the OLT and the PON is 32 users accessing the switch. The same congestion on the PON is occuring inside the Switch/Router to get to the GE link Network port of the switch/router.

With EPON or GPON the data rate to the home downstream is actually 1 Gig (not 100M as P-P switch/router at 100M bps). Based on time shared access not all 32 homes are getting data sent to them at the same time. Most internet traffic is bursty anyway since TCP/IP protocol is a Ping/Pong protocol which waits for a response back before sending more. Additionaly, the home users are sometimes sending information upstream which does not interfer with the downstream rates. The result is a PON is very efficient use of internet traffic patterns. The only real service that causes congestion is Video streaming (IP video). But if this is Broadcast video (network television) then there is no real problem with it either if a GPON or EPON protocol is used with IP Multicast. All ONTs then can receive one video stream on the PON and all home data ports can get copies of the video to the TV/PCs via IP Mulicasting within the ONT (similar to how coax cables and set top boxes do for Video today. Video to your home is coming on a Passive Network with share transport (Cable Modems act like ONTs,their bandwidth is shared on the coax), coax within your home.

The downstream traffic flow will thus accomplish sending data to each home at 100Mbps almost all the time if the data port at the ONT is provisioned to support that rate (100 Mbps burst or best efort rate set higher than guaranteed rate 30 Mbps).

The upstream traffic is likewise increased to more than 30 Mbps per home by taking advantage that not all homes are sending traffic upstream at the same time. One ONT has Gig speed upstream transmission for as long as no other ONTs are trying to send anything upstream; or share the Gig speed upstream with 2-10 other ONTs if that is all that are trying to send upstream information at any time frame - buffering in the ONT allows this to send at much more than 100M bps from each ONT (70-90% througput rates of the GE upstream direction is acheivable). With Dynamic Banwidth Allocation software the upstream traffic can also acheive 100Mpbs per second for any home data port nearly all the time.

To upgrade the network the ONT and OLT do not need to be replaced, split the PON into 2 16-port PONS and the bandwidth is doubled for guaranteed traffic and if the ONT is made right it can handle more than one 100M bps data port so that it can deliver more than 100M burst traffic to more than one device in the home. PON conversion for 32 to 16 ONTs can be planned for in the early fiber network and be implemented only on those PONs where bandwidth limits are being reached, not all PONs might need this.

PON architecture can also be encrypted so that no ONU can decrypt the other ONU traffic. There are new encyption methods that make it very difficult to break since the ONT and OLT ends pass partial encryption key information back and forth to change encryption for each ONT. Some methods make this encyption byte size very large to prevent breaking it before the OLT and ONT change the keys again, milli seconds between key changes.

All the P-P fiber designs cost too much compared to PONs currently for NA single family homes and they are more limited in burst bandwidth delivery (100M bps maximum). Gig P-P is much too expensive electronically. Even internationally PONs are being used for single family homes (about 20% of the homes). PON is not just a NA market application.

This is only one of many products that will help get the higher speed access to home users.

Integrator
Road Trip 12/4/2012 | 11:18:56 PM
re: FTTP Bidders Slashing Prices? It's called ITU G.983. Read it.
rs50terra 12/4/2012 | 11:18:56 PM
re: FTTP Bidders Slashing Prices? >"... Because of the high launch power, some other effects can creep in which makes it advantageous to split as close to the headend as possible."

This actually points to one of the biggest questions regarding PON. The advantage of PON over a simple Point-to-Point architecture is in the savings achieved by deploying less fiber and sharing one downstream tranmitter for 32-64 subscribers.
If the splitter is installed close to the CO, the fiber saving advantage is eliminated.
Now, the only selling point is in the electronics and the optics. However, a P-t-P system will use lower data rates and simpler receivers, so the economics are not clear. In addition to that, using a P-t-P architecture allows SPs to use 'traditional' L2/L3 switches using IP rather than ATM and thus achieving lower OpEx and CapEx.
Long Live Metro_Ethernet1 :-)
gea 12/4/2012 | 11:18:57 PM
re: FTTP Bidders Slashing Prices? "It uses SONET bit rates (155 or 622 Mbps) but that is the only thing that is close to Sonet. There is absolutely no framing in APON that even starts to resemble SONET."

Really? Are you absolutely sure? I was told that it was ATM-over-STS-3c-mapped OC-3 upstream, and ATM-over-STS-12c-mapped OC-12 downstream, and by someone who had been on the PON comittees from the very beginning. This would at least allow PON companies to leverages the existing framer chips rather than completely start from scratch.

If you know otherwise I'd love to see the standards document that proves it.

doco 12/4/2012 | 11:18:57 PM
re: FTTP Bidders Slashing Prices?
Very good description gea - especially the problems with a near 0 being stronger than a far away 1.

However - APON is not ATM over Sonet. It uses SONET bit rates (155 or 622 Mbps) but that is the only thing that is close to Sonet. There is absolutely no framing in APON that even starts to resemble SONET.

Of course the EPON people talk about leveraging Ethernet's volumes - BUT even with EPON you still have this near 0 being stronger than a far off 1. That is a problem that no other Ethernet phy has to deal with so EPON and APON sit in the same boat there.

gea 12/4/2012 | 11:19:00 PM
re: FTTP Bidders Slashing Prices? "Haven't been following too closely, but PON is already deployed to the curb anyway. It's called HFC."

Nope. PON is a specific architecture whereby a passive optical splitter is placed somewhere outside. A single fiber goes in to the splitter, and that optical splitter is split N ways downstream. That's not the bad part (though there may be some security issues as others point out).

The wierd part is upstream. Each enduser gets a tiny time window into which to send upstream info. ALl of the users' info gets precisely timed so that a full SONET frame is "glued" back together for the PON terminal to see (for those who don't understand, PON leverages ATM-over-SONET).

The problem is that the PON Rx is a very unusual beast, and upgrading its speed needs new physics for the odd way it must detect signal (ie, a nearby '0' can be stronger than a far away '1'.)

Upgrading the PON's speed will therefore be very difficult, and the PON Rx can not leverage price curves in either the Ethernet nor SONET realms.

However, when there's no alternative (ie, can't lay enough fiber), then in that niche it dominates. Outside that niche operators will avoid it like the plague.

However, HFC it certainly isn't.
whyiswhy 12/4/2012 | 11:19:01 PM
re: FTTP Bidders Slashing Prices? Haven't been following too closely, but PON is already deployed to the curb anyway. It's called HFC. The MSOs used to just use fiber for crossing long spans, but now they are going right to the last distribution box on the block.

The question is: what benefit would the phone company get from copying the cable company lead to the home? It's a huge me-too-play, leading only to price war. That's a bad play from a pure business perspective.

Their (the phone companies) play should be to wireless broadband, i.e. portability and mobility. Give me heads up display of traffic, news, etc while in the car. Movies on my cellphone, etc.

JMHO

-Why
optical Mike 12/4/2012 | 11:19:01 PM
re: FTTP Bidders Slashing Prices? The question is: what benefit would the phone company get from copying the cable company lead to the home? It's a huge me-too-play, leading only to price war. That's a bad play from a pure business perspective.


The answer would be to stay in business, cable modem subscribers are presently signing up faster than DSL. Once connected if service is satisfactory the only reason to change providers would be lower price or more/better offerings such as higher bandwidth to take advantage of new technologies like HDTV/SDV streams running at 10-20Meg, 100's+ digital cable channels interactive online gaming.
The cable companies are now picking off voice customers and the phone companies are losing revenue, they either can sit idly by or they can take an active roll in protecting their business.
data_guy 12/4/2012 | 11:19:03 PM
re: FTTP Bidders Slashing Prices? Yeah, 30 Mbits/s to my home absolutely sucks and is no good for anything.
I'm pretty neutral on the subject, but other than puling and whining, what do you have to say on the subject that's insightful, informative, and well thought-out?
The attraction to pon is that there are not a lot of fibers leaving the CO, and there's no active element in the field except at the users house. The only other way to do this is DWDM, which is too pricey at this point (expensive optics, and 32x more transmitters and receivers at the CO).
optodunce 12/4/2012 | 11:19:03 PM
re: FTTP Bidders Slashing Prices? Gea, I apologize for the forgotten "r"...I don't wish anybody to think I was Itallian...not that there's anything wrong with that!
optodunce 12/4/2012 | 11:19:04 PM
re: FTTP Bidders Slashing Prices? Gea, ZI always respect you point of view, and I also agree with you amendment...

And if you accept that premise, then PON will never be anything more than a nitch market product...

TO that end, there has to be more advance thought on a wide spread fiber system.

Does anybody know anything on Essex Corporation...www.essexcorp.com!

I would be interested in hearing any info other than wants on their web site propaganda!

gea Thanks for your input
photonsu 12/4/2012 | 11:19:04 PM
re: FTTP Bidders Slashing Prices? "PON is useful where you have a group of residences in a fiber constrained area. Other than that, it won't get used."

Unless you put a big-fat switch at the curb!
*******

FTTC/FTTH is THEE best solution for most applications! Why?

FTTC node is either network powered or from subscribers via copper power pairs in the fiber drop cable. This single point of network electronics is much more cost effective than hanging N boxes.

FTTC node is fed with GbE, OC-48, 10GbE. Node provides subscriber rate speed matching, remote provisioning and security. Subscribers can be fed with Fast Ethernet or OC-3 or GbE as required.
With PON, security will always be suspect no matter what the level of encryption. Just ask the White House if they would be happy to be fed off a PON that also feeds the French embassy!

With PON are the fibers leading to homes driven even when the subscriber isn't? What prevents him from eves-dropping? When a subscriber signs-up for the service, will a truck-roll be necessary to the passive-FTTC-splitter?

If I were Mr. Cisco, I'd be putting mini-routers all over the place (FTTC node). Fanning out with new MMF with options to use either 850nm or 1310nm VCSEL based drops to the home. When the customer is ready to take up the service he goes to Circuit City, Radio Shack, Comp USA or Good Guys and picks up the appropriate interface.

QED
gea 12/4/2012 | 11:19:05 PM
re: FTTP Bidders Slashing Prices? I basically agree with your assessment.

Basically, due to the inherent limitations of PON (including the cost most of the time), it won't get deployed unless it really has to. In urban areas I see no reason anyone would deploy PON instead of, for instance, using 1/10GbE and putting a big fat switch in the basement of MDUs.

However, as for the cost of fiber, I think you've missed a point there. These days there are a lot of choices for the enduser, so that places a limit on how long you can expect them to remain customers. Therefore, you can only spend so much and expect to get a return. PON is useful where you have a group of residences in a fiber constrained area. Other than that, it won't get used.

In other words, PON sucks excepts in the few cases where it's sessential. But cost does matter.
gea 12/4/2012 | 11:19:06 PM
re: FTTP Bidders Slashing Prices? I basically agree with your assessment.

Basically, due to the inherent limitations of PON (including the cost most of the time), it won't get deployed unless it really has to. In urban areas I see no reason anyone would deploy PON instead of, for instance, using 1/10GbE and putting a big fat switch in the basement of MDUs.

However, as for the cost of fiber, I think you've missed a point there. These days there are a lot of choices for the enduser, so that places a limit on how long you can expect them to remain customers. Therefore, you can only spend so much and expect to get a return. PON is useful where you have a group of residences in a fiber constrained area. Other than that, it won't get used.

In other words, PON sucks excepts in the few cases where it's sessential. But cost does matter.
Curious George 12/4/2012 | 11:19:06 PM
re: FTTP Bidders Slashing Prices? Thanks for the standards correction.

>>>It seems that most ONT's operate at under 7W at 48V DC, so what you suggest is a neat idea. I don't think anyone has implemented it. <<<

7 watts for an ONT would not be considered typical but rather on the low end. Actual devices under test today seem closer to 10 W, but still within the powering spec over cat 5.

>>>I'm also intrigued by the "collar" you mention. Whilst it is conceivable to tap power by induction in this manner, it would have to be performed after the electical meter. This means that in theory one still has to deliver the power cable through the wall.<<<

The collar is really not an inductive device, but rather a pluggable cylinder that has connectors top and bottom which sits between utility cabinet and meter. Externally all you see is the outer circumference of the device - hence the name.

>>>Actually most SP's that have researched this are happier to use sealed lead acid batteries than Ni-Cd or Ni-MH. They respond better to a long shelf life of trickle charging, typically lasting 10 years, and are widely deployed today inside emergency lights, and as backup power for Alarm Systems. <<<
Disagree, most SP I have spoken with are leaning toward "standard" batteries ("D", "C" NiMH, etc.) that may not perform as well or even last as long as "sealed" wet-type bateries, but rather are available and replaceable by the end user.




optodunce 12/4/2012 | 11:19:07 PM
re: FTTP Bidders Slashing Prices? Of those who are in the know...do you not see the obvious flaws in the PON design concept...no wonder why there is not wide spread compliance and deployment for this technology...as a shared network with bottleneck collision points upstream and braodcast information downstream this is a very unsecure network with little reduncancy and very little future proofing...

Why do we still continue like lemings implement inferior designs as stop gap measures for building the most important infrastructure our communication pathways.

WHy is everyone convinced that the RBOC's and others will only deploy fiber if it cost less than deploying copper! If you have 10000 times greater performance, reliability and longevity, than any good business would implement that technology even at a premium because in the long run OP EX will be significantly less and services and revenue will be much greater through value added services...

THE PROBLEM IS THAT PON SUCKS AS A SOLUTION FOR FIBER TO THE HOME! AND INSTEAD OF PROVIDING A BETTER DESIGN WE ARE TRYING TO REDUCE THE COST OF A BAD DESIGN TO SHOVE IT DOWN THE THROATS OF THE CARRIERS!

LASTLY, What drives me nuts is that PON is used interchangably for FTTH or P! It is not the only fiber to the home technology available.



Rupert_1 12/4/2012 | 11:19:09 PM
re: FTTP Bidders Slashing Prices? On the subject of powering, is anyone up-to-date on the state of photovoltaics? At some point that may be the ideal solution. A quick search comes up with this from a company called PhotonicPower:

"InP Photovoltaic Power Converter (PPC-LW)

Photonic Power is pleased to announce its new InP photovoltaic power converters. The InP-based device can operate in the wavelength range from 1310nm to 1550nm. This allows the transfer of power over fiber at distances up to 20km or more. Consequently, remote antennas can be accessed without any need for electrical wiring.

With its launch of the InP power converters, Photonic Power is entering new exciting Communications areas where applications such as conversion from analog to digital TV and powering wireless antennas are increasingly using optically powered technology.



The question is, how much power is delivered? I seem to recall devices at lower wavelengths delivering several watts, several years ago, but they weren't yet cost-effective. This doesn't seem to be a real active R and D area for some reason. Can anyone add any updated info on photovoltaics?
wap545 12/4/2012 | 11:19:09 PM
re: FTTP Bidders Slashing Prices? AFC had no choice but to win here if they were to remain a viable vendor in the Telecom Space. If the Motorola/Quantum Bridge or Alcatel team had won this RFP AFC DLC business would go away over the next 2-3 years.
Verizon gave them the business due to Low Ball pricing (plus I feel Verizon is in no hurry to deploy FTTH and the other two vendors teams can deliver a viable FTTH solution today) and also they will allow them (AFC) 14 Months to develop and demo a viable FTTH solution.
What will be really interesting to see is what Vendor these RBOC will actually put in field trials in 2004.
My bet is that Alcatel and Motorola/Quantum Bridge will get some field time in both the RBOC and the IOC space.

wap545 12/4/2012 | 11:19:09 PM
re: FTTP Bidders Slashing Prices? AFC had no choice but to win here if they were to remain a viable vendor in the Telecom Space. If the Motorola/Quantum Bridge or Alcatel team had won this RFP AFC DLC business would go away over the next 2-3 years.
Verizon gave them the business due to Low Ball pricing (plus I feel Verizon is in no hurry to deploy FTTH and the other two vendors teams can dliver a viable FTTH solution today) and also allowed them (AFC) 14 Months to develop and demo a viable FTTH solution.
What will be really interesting to see is what Vendor these RBOC will actually put in field trials in 2004.
My bet is that Alcatel and Motorola/Quantum Bridge will get some feild time in both the RBOC and the IOC space.

Ben Crosby 12/4/2012 | 11:19:13 PM
re: FTTP Bidders Slashing Prices? There's quite an old study of powering issues published by EURESCOM circa 1998. I'm not aware of anything more up to date, but would happily be pointed at something.

Correcting some factual errors.

>>>
A third and more elegant approach is to take advantage of standard 802.3ae - power over ethernet. In this case, a wireless 802.11x router supplies power from a wall socket through the cat 5 cable that connects to the ONU on the side of the house.
<<<

First, the standard for power over ethernet is actually 802.3af (not ae).

Second, The IEEE's 802.3af specification calls for power source equipment (PSE), which operates at 48 volts DC, to guarantee 12.95 watts of power over unshielded twisted-pair cable to data terminal equipment (DTE) 100 metres away.

It seems that most ONT's operate at under 7W at 48V DC, so what you suggest is a neat idea. I don't think anyone has implemented it.

I'm also intrigued by the "collar" you mention. Whilst it is conceivable to tap power by induction in this manner, it would have to be performed after the electical meter. This means that in theory one still has to deliver the power cable through the wall.

If you look at Cisco's typically carrier class approach, they have obviated the need for holes through walls with their 1xxx ONT products, suggesting instead that:

"The power cord coming from the power supply may run through a small hole drilled in the wall or out a slightly opened window."

http://www.cisco.com/en/US/pro...

I can't find any references to other vendor's power suggestions.

>>>
People want to stay away from wet cell batteries since they can leak and make a mess/cause corrosion or can poison a pet in the garage - usually placement.
<<<

Actually most SP's that have researched this are happier to use sealed lead acid batteries than Ni-Cd or Ni-MH. They respond better to a long shelf life of trickle charging, typically lasting 10 years, and are widely deployed today inside emergency lights, and as backup power for Alarm Systems.

>>>
Eventually, UPS may go away from a practical perspective since most people cannot use the phone if main power fails due to a predominance of wireless phones.
<<<

I really doubt this. After the North Eastern power outage, all my local technology stores (Futureshop, RadioShack, Best Buy - even Staples in this case) had sold out of UPS'. APC, Belkin, Energizer and others have all brought out sub $100 products that they target at end users. A surprising number of my neighbours have UPS' for their computers and also use them for the cordless basestation for the phone.

As you rightly said, it takes something without power to prompt people to understand the value of certain technologies.

In most cases, there's going to be holes drilled in walls. ;)

Ben.
Curious George 12/4/2012 | 11:19:15 PM
re: FTTP Bidders Slashing Prices? Power is a key issue. A couple of ways to handle. One is a "collar" placed on the power meter which taps appropriate voltage/current and also sends data through the PON to the power company with your usage stats. In some places, this is valuable enough to the utility that they will pay for the collar. Another potential method is drilling a hole (roughly $100) through the exterior wall and tapping into home power. A third and more elegant approach is to take advantage of standard 802.3ae - power over ethernet. In this case, a wireless 802.11x router supplies power from a wall socket through the cat 5 cable that connects to the ONU on the side of the house. Drilling holes for data cables is cheaper, smaller hole less danger of hitting a pipe, easier to seal, and also low voltage. In all cases batteries are used to provide some form of UPS. This is another issue. What is being proposed is that the RBOCS would supply the first set of common batteries (D's for example) which are trickle charged. When end of battery life is approached, a signal is sent from ONU to the OLT or headend. This then triggers an email message to you indicating you should change your batteries. If you don't you take similar risk as a smoke alarm - your problem. People want to stay away from wet cell batteries since they can leak and make a mess/cause corrosion or can poison a pet in the garage - usually placement.
Eventually, UPS may go away from a practical perspective since most people cannot use the phone if main power fails due to a predominance of wireless phones. It is hard to find a RJ11 powered phone these days. However, it will only take one publicized case of a baby born on the front lawn due to an emergency during a power failure that will trigger an extension for this powering requirement.
Curious George 12/4/2012 | 11:19:15 PM
re: FTTP Bidders Slashing Prices? A bit more detail as I understand it: PON first then AON.

PON. FSAN standards with current implementation being BPON - what the RBOCS asked for. BPON uses three wavelengths. 1550 overlay for standard video broadcast that everyone sees. 1490 downstream from the headend for voice and data - also everyone sees. 1310 upstream for voice and data from the home. Downstream data at 622 Mb/s is read by the intended ONT on the side of the house as per the PON MAC stack which takes care of addressing, but again, everyone sees. The downstream path is relatively simple - upstream is another story. Upstream is handled in a burst mode at 155 Mb/s. Essentially when the PON is setup, the OLT or head end does something called ranging. Address pings are sent to individual ONT's (homes) and time of flight return pings clocked. This time of flight is used to determine when the OLT pays attention to a bit stream coming from a particular home and is used to control flow of traffic. The transit time along with other addressing information is used to ensure the appropriate bits have originated from the intended ONT or home and responses do not contend with each other in the shared upsteam pipe (remember, all 32 go into one in the headend). The OLT or headend is in complete control. BPON frames are ATM-based, hence the familiarity for the RBOCS. The next incarnation of PON is GPON, which allows for higher transmission rates and introduces a new protocol called GFP - Generic Framing Protocol which allows for mixed ATM and IP payloads.
The value of PON as stated by the RBOCS is that there are absolutely NO ACTIVE ELEMENTS in the outside plant. External to a building, everything is passive. This explains the intended OpEx savings they anticipate and also is consistent with statements from Verizon like "we spend 3-4% of our total budget on Cu rehab, so roughly every 30 years we rebuild the network. Why would we build another Cu network, users will be migrated to fiber in any case and next gen access network will be fiber based, even if FTTH is slow to be broadly deployed"

AON: Active network is essentially IP based with all of the swithing, collision control, etc. that is imbedded in that protocol. However, unless point to point networks are used - basically a 1 or 2.5 Gb/s link from the head end to the basement of a large building, switching elements are found in the outside plant. For example one might have a GigE link to a neighborhood switch which then distributes and aggregates upstream traffic from separate drops, this is point to multipoint. There is an ethernet pon variant called EPON which attempts to get rid of the outside plant active elements, but typically with a hit in efficiency and overhead since switches are quite good at controlling and aggregating traffic. Also, IP payloads can be of varying length, unlike ATM so timing and control is a bit messier.

For ethernet stuff, lots of good reading on the EFM (ethernet for the first mile website). For PON stuff, you can read standards or get bootleg copy of RFP - in either case good bedtime stuff.

From an optical perspective, the data/enterprise space is driving GigE modules to relatively low price points so volume is consolidated between enterprise and access.
On the PON side, a triplexer is a very difficult part to build with a target of $50 and less in high volume. The trickiest bit is the linear PD and amplifier for the video overlay. The data streams are relatively straightforward with digital PD and standard TIA and a directly modulated FP. On the OLT side, the challenge is the 1490 launch power since it is split 32 ways. The 1550 signal is amplified by and EDFA so plenty of juice. Because of the high launch power, some other effects can creep in which makes it advantageous to split as close to the headend as possible.

Hopefully enough detail here to draw the conclusion that PON can and does work but is a North American phenomenon at present. Some real technical challenges here. Much of the rest of the world is entirely digital and IP based with the bulk of the broadband brought to the basement of an MDU (multi dwelling unit) and distributed via Cu within the building.
Ben Crosby 12/4/2012 | 11:19:17 PM
re: FTTP Bidders Slashing Prices? >>>
I am cusrious as to how carriers are going to power these devices that sit outside of the house in the access NID. If that is where the ONT will reside.
<<<

Since you have to run copper through the wall to the inside anyway, the powering is typically from inside the premise. To meet lifeline requirements many SP's are advised to deploy a device such as APC's powershield.

http://www.apc.com/products/fa...

Cheers,
Ben.
network 12/4/2012 | 11:19:18 PM
re: FTTP Bidders Slashing Prices? The PON solutions I have seen have a very nice triple play soluton. I am cusrious as to how carriers are going to power these devices that sit outside of the house in the access NID. If that is where the ONT will reside.

The demarcation point is of great interest especially as regulatory bodies weigh in on the bundling of services.


dodo 12/4/2012 | 11:19:30 PM
re: FTTP Bidders Slashing Prices? Opto dunce

This is my understanding:
There are 2 types of PONs: APON and EPON
The OLT send frames of data ( ATM or Ethernet) down to the 1:32 splitter.
The splitter sends same set of frames to each ONU.The ONU filters out only the frames that are specific to a user and then discard the rest ( which has not been encoded that said user address)
The bandwidth is shared but not the data.

This is my limited functional view- been away from BB access for a while.

Hopefully someone can give us more info
optodunce 12/4/2012 | 11:19:30 PM
re: FTTP Bidders Slashing Prices? If understand the PON design there is 1 fiber split down to serve 32 subscribers. Up-stream will be using one channel and down stream another and broadcast a third.

How does the information get isolated to one 1 of the 32 users...in other words doesn't all 32 users have access to the information that is assigned for user 1?

How can this be carrier class type service? can't any of the 32 users tap in on the information?

I am a bit confused on how effective PON design will be in the log run...
Curious George 12/4/2012 | 11:19:32 PM
re: FTTP Bidders Slashing Prices? Pricing is $350 for ONT and $50 for associated OLT. RBOC's have stated that this pricing "will work". Volumes were listed at three levels in the RFP, all rather high and intended to stimulate cut throat pricing. Most likely '04 will remain in the trial mode with deployments in the 10's of thousands, '05 a factor of 10 above that and significant volume >1M not until '06. With regard to market segments, the Greenfield builds will be dominated by fiber terminations at the home. Given roughly 1.4M new homes/yr, a number of at least 1M/yr is reasonable, beyond that we must include overbuilds, a much tougher financial hurdle. Also "take rates" for triple play service (voice, video, and data) will be significantly lower for overbuilds making the business case even more challenging. Total cost per home passed is roughly $1500 even at the 100K/yr run rate. Of this, roughly 40% is associated with construction labor that does not significantly benefit from increased volume. While any cost reduction is attractive, there is then a large portion that component suppliers can't really impact - except for designs that are inherently simpler to install. Other cost elements of the ONT like battery backup, SLIC and voice codec chips, power/protection circuitry, etc. are already rather cost reduced. The triplexer and processor block (some sort of CPU, PON MAC, flash memory, etc.) is about all we can significantly impact. Triplexers need to hit prices on the order of $60 in volume to be attractive.

Strategically, forward pricing is inevitable given the extremely competitive nature of today's PON market. The key issue is one of timing. ONT suppliers must work very hard between now and '06 to dramatically cost reduce and achieve a reasonable business model when volume deployments finally occur. It would actually be a major business problem if volume hit much sooner than '05 as the technology development and progress in the standards bodies would not have sufficient time to come to fruition. On the other hand, sub component suppliers cannot cost reduce until sufficient volume is available - this is the crux of the Catch 22 we are in. Government intervention could play a significant role in breaking the deadlock if we can make bringing bandwidth to the home a national issue that impacts economic recovery (e business, etc.) and education. Last year we ranked 4th in broadband delivery to the end user, this year we are 6th and falling further behind.
strands555 12/4/2012 | 11:19:35 PM
re: FTTP Bidders Slashing Prices? re: " cannot see why companies are being criticized for pricing their product on the assumption that the market is going to be successful. If the market is successful then they will make money at the price. Am I missing something here? "

It's not necessarily true they will make money (as in profits, not just revenue). Generally the early market's high margins are where you recover R&D expense. If you can pay those off before the margin erosion, then the late market at razor-thin margins and decent mkt share is an OK proposition for some (but not all) companies.

If the later market becomes hyper-competitive (like DOCSIS cable modems did), then even if you are an early market leader (like 3com was) the business returns just don't materialize. When 3com dropped out they were still #2 in market share.

It's not necessarily healthy for the overall market segment when one company "skips over" the high-margin period of R&D recovery to try and get an "edge." Often, everyone else will have to follow suit, and no one recovers R&D costs, or it is pushed out so far into the future that it's no longer a viable business to be in. So they do get an edge all right...they dive right onto the edge of the razor and bleed to death.
HiTekRedNeck 12/4/2012 | 11:19:35 PM
re: FTTP Bidders Slashing Prices? The price is sub $300 per home for ONT, OLT.

Predictions were it would start at $1,200, and RFP would drive price points to $500-$700. Yikes.
photoelectric 12/4/2012 | 11:19:36 PM
re: FTTP Bidders Slashing Prices? Re: "Do you think companies that give up margins early on in hopes of making it up on volume are acting wisely? Let me know what you think and why."

Clearly, one has to jumpstart this business and price for volume. Without making this investment, why bother trying to play in this market? The problem for start-ups is their ability to subsidize the early years.

I would guess Alcatel is the clear early leader.
DCITDave 12/4/2012 | 11:19:36 PM
re: FTTP Bidders Slashing Prices? It's a bit tough to generalize like that in this case. There have never been deployments of FTTP gear in the volume that this RFP is calling for, so whatever something cost before this RFP is more or less irrelevant.

We are working to get some idea of what prices are. We're trying to find out what's been bid and at what volume, but it's not easy. This is closely guarded stuff.

Even without hard numbers, it is worth opening a discussion of forward pricing and what its impact might be on the companies involved.

Do you think companies that give up margins early on in hopes of making it up on volume are acting wisely? Let me know what you think and why.

-- phil
optodunce 12/4/2012 | 11:19:37 PM
re: FTTP Bidders Slashing Prices? DA...this is business 101...in gross terms of course risk relative to the market portential and it is the bold who have disruptive technologies in a questionable market that often get ahead...but the consept of lost leading or front ending often finds failure for the company that ventures out...quite often it is the second and third players that pick up the pieces and learn from the inevitable mistakes of the adventurer that wins in the long run.

You cannot see why they are being criticized...well just look back at worldcom, global crossing, and others that had Op ex expenditures and cap ex that would take 20 lifetimes to pay back...
photonsu 12/4/2012 | 11:19:37 PM
re: FTTP Bidders Slashing Prices? What I find interesting is how the BOCs dictate the solution, and then complain about the price. Such is the lifecycle of PON. Gets down-right boring after a while.
One of these days maybe, just maybe they will state the problem and let all technologies compete as a solution. Until then .... yawn!!
dljvjbsl 12/4/2012 | 11:19:38 PM
re: FTTP Bidders Slashing Prices? I cannot see why companies are being criticized for pricing their product on the assumption that the market is going to be successful. If the market is successful then they will make money at the price. if the market is unsuccessful then they will go broke no matter what they price the product at.

Am I missing something here?

The Ottawa Citizen contained a 3 full page article on Scott Marshall and Ceyba. In it, Marshall was strongly criticized for building Ceyba to deliver a prodict in late 2002. Some of the founders and the others said that Marshall should have scaled back the company even more dramtically than he did so that the company could last through 2003.

I was puzzled by this. If the ULH market had existed as the founders hoped, Ceyba would have had a product to deliver to it. Since the ULH market did not appear then the Ceyba bankruptcy was inevitable and it would not matter that the company could have extended its death process a little bit longer.

Amrshall built a company on a strategy that had a chance of winning. The foundes and the Ottawa Citizen adivcated a strategy that could only fail.

The same analysis pertains to this issue of pricing. One must assume that one can win. Strategies that are based on the assumption of failure can only fail.
optodunce 12/4/2012 | 11:19:38 PM
re: FTTP Bidders Slashing Prices? Phil,

Your article header is FTTP Bidders Slashing Prices? Yet no where in your article do you reference what the baseline price is! YOu do reference that no one will go on record, but so that we can recognize the relative impact of your article can you at least give some estimated references in line with what the cost breakdown is regarding FTTP.

Quite often LR pens commentary on subjects that are loose and incomplete and do not give even the educated reader frame of reference.

I think it would be quite simple with relatively little work to breakdown the estimated costs per drop (i.e. FTTP equipment costs of $2,200 per drop for a fully populated rack of 1000 subscribers, installation costs of about $ 1,100 and cable costs....)

Thanks
gpiccirilli 12/5/2012 | 1:04:44 AM
re: FTTP Bidders Slashing Prices? Hello,

Do you know any vendors that do PON over MMF? I have MMF in the last mile and would be interested to see if there is a chance in using it to deliver high speed services. Thanks and regards.


Giovanni
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE