ajschmitt
join:2005-11-30
| My Take on Lightspeed
I'm putting this in a letter to AT&T to see how they respond. I think they have under-reached with Lightspeed and are going to have to do everything again 10 years (or less) down the road.
All of the recent defensive comments from AT&T would indicate to me they are trying to hide the fact this rollout just isn't going to be competitive.
Link: AT&T Lightspeed and the Jedi Mind Trick |
|
ronpin
Imagine Reality
join:2002-12-06
Nirvana
 ·AT&T Southwest
| I was all over this a year ago -- I finally gave up.
If SBC is willing to ignore the 6 mile range and "no field electronics" features of passive Fiber to the Premise (FTTP) -- then optical OLT's can be crammed into the new "Mini RTs" for LightSpeed -- and homerun fiber ran from there (an ugly contradiction to FTTP purists). That would mean higher OSP costs for eternity.
Still, the $10/ft cost for directional drilling of buried conduit is the real sticking point. Most of the affluent subdivisions -- where a fast ROI can be had -- have no aerial wires. So a typical 75 ft. lot frontage costs $750 -- just for the conduit passing. The home ONT cost is down to $300-$400 and STBs are maybe $100.
Verizon clearly saw the advantages of doing it right the first time. My years in software taught me that most companies prefer to do it wrong the first time -- if that'll get things started. They can count the bodies later.
Lets hope a cheap non-invasive method of conduit placement can be developed.
--
"Now is the time for all good men to come to the aid of their country" - and stop the NeoCons |
|
wierdo
join:2001-02-16
Tulsa, OK
 ·Future Nine Corpor..
·Teliax VOIP
| reply to ajschmitt
Two or three words, depending on how you look at it..RF interferers.
BTW, I saw one of SBC's contractors out laying fiber a couple of years ago with one of those vibrating plows, and I couldn't discern any damage to the grass after a day or two. Of course, they still had to bore under the streets, but it's not necessary to bore the entire length of the conduit. The only people doing that are those with more money than sense, which would be SBC after hiring Orius, who bored quite literally 4 miles for SBC late last year, despite there being no pressing need to bore in that location rather than plow.
5 years ago, they just said "screw 'em" and used a trench digger and laid it in, but they seem to not like to do that these days. |
|
ronpin
Imagine Reality
join:2002-12-06
Nirvana
·AT&T Southwest
1 edit | said by wierdo  :... I saw one of SBC's contractors out laying fiber a couple of years ago with one of those vibrating plows, and I couldn't discern any damage to the grass after a day or two. Of course, they still had to bore under the streets, but it's not necessary to bore the entire length of the conduit. Home sprinkler systems preclude the elegant vibrating plow from a lateral crossing of suburban frontages (drat!).
There are cheaper ways to run "conduit" -- but they're too "out-of-the-box" for a megalith to consider (see picture) -- yet another reason why we'll be stuck with copper for the foreseeable future (another radical idea-> »FiOS's one mistake? )
--
"Now is the time for all good men to come to the aid of their country" - and stop the NeoCons |
|
wierdo
join:2001-02-16
Tulsa, OK
·Future Nine Corpor..
·Teliax VOIP
| Very few people around here even have sprinkler systems, and those that do need to keep them out of the utility RoW 
BTW, what's the going rate on a GPON splitter these days, if you know?
I'm still in favor of the active electronics in the field capable of bringing 400Mbps unshared to each home, but that's mainly because I'm a bandwidth whore and a firm believer that this whole asymmetric speed thing is bad for everyone. Do a little uploading, and unless you own a router doing QoS, your download speed suffers immensely due to the increased RTT.
Of course, that architecture isn't cheap..I priced it out at about 16MM to serve the entirety of a spread out city of a little under 70,000 including fiber, cabinets with A/C, big switches, fiber converters, an IPTV headend, and STBs. That said, it is quite literally scalable up to the current limits of fiber technology, being able to bring 10Gig circuits across town with ease.
The only problem is the same that SBC has and will be continuing..after the power is out a couple of hours, there goes the cable and internet. It's hard to cram all the equipment into 22U, much less battery plant that will last more than 4 hours, after all. 
That said, going with PON right now isn't at all stupid, as the standards are constantly evolving, and it probably won't be long until they get OC-192 on the downstream and OC-48 on the upstream, and anyone with a PON network can just change out a little hardware and be operating at those speeds, even without a change of existing customer hardware if they use WDM. I just have a philisophical problem with providing what I consider to be substandard asymmetric service, so if I was going to do such a thing, I'd do it right, since it'd be profitable doing it "properly" anyway.
With bandwidth prices what they are, if you can cut out the local loop, it's not even unaffordable to provide 10Mbps symmetrical at $40/mo or less for a small outfit, and it's easily affordable for someone doing a deployment on the scale of Verizon or SBC.
Back down to planet Earth, I think it's immensely stupid to buy a bunch of active electronics you're going to can shortly anyway. VDSL isn't going anywhere. It turns out I'm a mere 1kft from an RT, and the best I can get is around 4Mbps thanks to interference. If people have the usual DSL problems with their TV service, they'll drop it like a hot potato. People really have a problem when their DVR doesn't record their show because the "cable" was out. |
|
nunya
SEE ROCK CITY 475 MILES
Premium,MVM
join:2000-12-23
O Fallon, MO
clubs:
·AT&T CallVantage
| reply to ajschmitt
FWIW, I'm sure the honchos are already aware of this and have been advised countless times. As much as we like to make fun, these are not stupid people.
The article above does have some inconsistencies and is slightly misleading. For example, conventional DSL will not be the last mile solution.
That being said, 25 - 35 Mbps is not going to cut it. Not now, and especially in the future. The last mile cable that is being "conditioned" for FTTN is in varying states of decline. The amounts of money being poured on it would be better spent on a steady FTTP push. Reconditioning the existing plant is turning into a money pit, with few benefits. The only good thing about FTTN is that the fiber will be that much closer to the end user.
There will be a mad dash to push FTTP once the FTTN proves that it will not be able to satisfy consumer product demand.
--
Nevermind the bullet with your name on it, try to avoid the shrapnel addressed to occupant. |
|
ronpin
Imagine Reality
join:2002-12-06
Nirvana
·AT&T Southwest
4 edits | Passive FTTP will be obsolete for a future FTTN upgrade. You'll recall FTTP was designed to handle voice, video and data -- all on separate lasers (1310nm, 1490nm, 1550nm). Also, the passive feature allows a 6 mile range, no field electronics and 1x32 passive splits per fiber (to reduce trunk fiber counts)
None of those features are needed to upgrade LightSpeed. The FTTN architecture already combines video and data onto the same pipe. Voice/VoIP is also just data.
The close-in FTTN nodes (less than 5000 feet) remove the need to reduce fiber trunk strand counts and thus the need to share fibers via 1x32 passive splits per fiber. Why? In the FiOS "centralized splitter" implementation -- a passive splitter pedestal is placed within 5000 feet of homes. Individual strands are then "homerun" to each house -- via high-count ribbon cables. This is precisely analogous to running high-count ribbon cables from the FTTN nodes to homes -- to replace the current copper pairs.
So, if you followed that terse explanation -- no shared/split fiber or multi-laser (WDM) "FTTP" will be needed at all. What remains is actually a classic "active" fiber distributed "star" network. Fiber -- in this case -- adds [only] OCxxx or gigE speeds for HDTV and data applications. The distance feature of fiber will not be needed to overlay an old close-in FTTN network.
Get it? an actual FTTP overlay to replace a FTTN network would be silly. An "active fiber" network can precisely duplicate the old FTTN network -- only with virtually unlimited future capacity. Cheap "media converters" can even maintain a classic copper signal at each homesite if desired (300' max range) One advantage of active fiber is "symmetrical upload speeds" will be available. FTTP has asymmetrical upload speeds due to the need to share upload time slots on the shared fiber. No problem on "active fiber". So SBC will wind up with a "better" fiber network than Verizon's FiOS (unless FiOS switches to uDWDM VC's -- unique lambda's for each user on a currently shared 32 user fiber strand -- a unique virtual fiber for each user).
--
"Now is the time for all good men to come to the aid of their country" - and stop the NeoCons |
|
nunya
SEE ROCK CITY 475 MILES
Premium,MVM
join:2000-12-23
O Fallon, MO
clubs:
·AT&T CallVantage
| reply to ajschmitt
The existing plant is in no condition to meet the 25 Mbps that AT&T says it will need. You need perfect cable. It aint perfect. Far from it. It's chewed up, taped up, tied up, pieced together, "make it touch so it'll talk" stuff. Schlepping 1.5Mbps DSL across a pair is one thing. Trying to push 25Mbps? I'm very concerned. The amount of money being spent to "recondition" this stuff could be invested in new fiber. I've seen over $110,000 spent on a few blocks of 40 - 50 yr old plant. After over runs, I'm guessing $10,000 - $20,000 more.
Call me crazy, but I think that would be better spent on an upgrade rather than a re-grade.
FWIW, the existing plant is only being conditioned to 3kft from the node. Anyone past this point (even one span) is left off.
--
Nevermind the bullet with your name on it, try to avoid the shrapnel addressed to occupant. |
|
ajschmitt
join:2005-11-30 | reply to ajschmitt
Thanks to everyone for the comments and info. The level of technical expertise on this board is impressive. |
|
ajschmitt
join:2005-11-30
| reply to ronpin
One thing you are missing - VZ runs a a multiple fiber bundle (like 40) to the splitter cabinet. Even though one of those fibers are being used, nothing precludes removing the splitter and then wiring up the home runs 1:1. At the very least, it would be easy to further partition the network to reduce the number of users per fiber.
And if none of this works, there is always WDM-PON, though I think it is a technology in search of an application. |
|
ronpin
Imagine Reality
join:2002-12-06
Nirvana
·AT&T Southwest
| said by ajschmitt :One thing you are missing - VZ runs a a multiple fiber bundle (like 40) to the splitter cabinet. Even though one of those fibers are being used, nothing precludes removing the splitter and then wiring up the home runs 1:1. At the very least, it would be easy to further partition the network to reduce the number of users per fiber. And if none of this works, there is always WDM-PON, though I think it is a technology in search of an application. Each of those 40 fibers from the Verizon CO to the splitter pedestal are terminated in optical OLT ports in the CO that mux up to 32 users each. So those 40 strands will serve up to (40x32) 1280 homes -- near the pedestal.
WDM-PON is actually the term being used to describe the current 3 laser BPON system. "DWDM" (Dense Wavelength Division Multiplexing) is "in search of an application" -- in the Last Mile. The ITU "color grid" defines 96 wavelengths in the 1500nm "C" band for DWDM . Really, only 32 wavelengths are needed to give each current FiOS user a unique "virtual fiber". I did err earlier (above) when I cited uDWDM as the method needed to do that (ultra-DWDM defines 1024 lambdas with much tighter spacing).
--
"Now is the time for all good men to come to the aid of their country" - and stop the NeoCons |
|
ronpin
Imagine Reality
join:2002-12-06
Nirvana
·AT&T Southwest
| reply to nunya
said by nunya :The existing plant is in no condition to meet the 25 Mbps that AT&T says it will need. You need perfect cable. It aint perfect....Call me crazy, but I think that would be better spent on an upgrade rather than a re-grade. No argument here -- but the fact is "big-mo" is in-play (and SBC forgot to hire me). FTTN is the name of the game for-now. Fortunately most of the suburban neighborhoods that will likely first see FTTN have been built in the last 20 years -- and have relatively good F2 pairs and terminals. The current round of line-conditioning at least proves that SBC has their eyes "wide-open" as to what they're getting into.
Sure we all wish SBC would do what Verizon is doing with FiOS -- but I'm resigned to the facts -- and the improvements that FTTN will bring. The cablecos need some heat -- and LightSpeed will do that.
My earlier technical treatise on a future fiber upgrade for FTTN only means I accept the years it will take to get there and the architecture applicable to an FTTN upgrade at that time. We can always hope for a fundamental breakthrough in physics in the meantime  )
Remember, buried conduit costs are the real enemy here -- not SBC. I'm hoping that the concept of a "pilot mole" takes-off and solves the cost problem. The pilot mole concept involves pushing a tiny projectile through soil -- under streets, sidewalks and driveways. The resulting hole is just big enough to drag a thick pull-wire in the reverse direction. That cable is attached to a 2" slug mole that is then dragged-back, compressing the soil, on-return, for conduit placement, while being reeled back to the pilot mole site. The holdup is that current "pilot moles" aren't steerable -- like directional drills. It'll be much cheaper than the traditional [$10/ft] directional drilling that Verizon is doing now.
--
"Now is the time for all good men to come to the aid of their country" - and stop the NeoCons |
|
ajschmitt
join:2005-11-30
| reply to ronpin
said by ronpin :Each of those 40 fibers from the Verizon CO to the splitter pedestal are terminated in optical OLT ports in the CO that mux up to 32 users each. So those 40 strands will serve up to (40x32) 1280 homes -- near the pedestal. Not according to the VZ lineman I spoke to. He said the cable that went from the splitter box (which serves maybe 100 homes) is practically unused.... that they just pull fat fiber. I wish I knew exact fiber counts, but this would mean that only 10% of the fiber is lit, and many extra OLT ports could be added in the future. Or, some users could be moved to 1:1 connections.
said by ronpin :WDM-PON is actually the term being used to describe the current 3 laser BPON system. "DWDM" (Dense Wavelength Division Multiplexing) is "in search of an application" -- in the Last Mile. The ITU "color grid" defines 96 wavelengths in the 1500nm "C" band for DWDM . Really, only 32 wavelengths are needed to give each current FiOS user a unique "virtual fiber". I did err earlier (above) when I cited uDWDM as the method needed to do that (ultra-DWDM defines 1024 lambdas with much tighter spacing). I think we are splitting hairs on terminology. I'm talking about what was announced by Korea Telecom and enabled by Novera Optics
»noveraoptics.com/htmls/company_overview.html |
|
d_l
Barsoom
Premium,MVM
join:2002-12-08
Reno, NV | reply to ronpin
Directional drilling may have to be reserved for rocky ground. It is overkill for areas that are just dirt. |
|
wierdo
join:2001-02-16
Tulsa, OK
·Future Nine Corpor..
·Teliax VOIP
| reply to ajschmitt
I disagree that Lightspeed will put any pressure on cable companies, except at the very low end, which is where it does most people (and the companies) no good. There's simply not enough bandwidth available to make it truly useful. Call me when they figure out how to do 2 dual tuner HD-DVRs while still having bandwidth left over for a 3rd (and 4th, in many homes) SD stream, while still pushing 6/1 worth of Internet service over the pipe.
Cox responded to Lightspeed in most of their areas months ago by bumping IP bandwidth up to 9/1; only a few areas formerly serviced by TCA that they haven't yet sold are still at 5/768. Cable has never had a technical problem delivering any reasonable number of HD and SD streams in any combination, nor do they have a bandwidth limitation preventing them from pumping up the speed to FIOS-type levels.
The only people this will put pressure on are the people who decided it was a good idea. The only way their bacon will be saved is if GPON turns out not to be beefy enough to do everything they want and they are forced to go with what ronpin was calling an active star fiber network due to bandwidth requirements. As he also pointed out, though, DWDM on the upstream can be used to easily scale a PON network to speeds equivalent to individual strands to each home, and will probably be cheap enough by the time BPON or GPON with a seperate cable wavelength will run out of room.
If Lightspeed had enough bandwidth behind it to actually force the cable companies to compete in any meaningful way, it wouldn't really matter to me that it was not the best technical solution possible, but it's not. It might be enough for this year, but next? They've got $500 or less HDTVs at Wal-Mart now. How long do you think it'll be before people want dual tuner HD DVRs for all of their HDTVs like they do for their SDTVs? Oh wait..that already happened. 
AT&T is deploying a solution for the year 2004, or stretching it a bit, 2005. In 2006, Lightspeed will not cut it. If it does, it'll only be by the barest of threads that it can hang on, one of which will be unacceptably high compression ala DirecTV to make 4xHDTV possible. It remains to be seen, but as I was mentioning earlier, I doubt channel bonding will end up being a real solution. |
|
ronpin
Imagine Reality
join:2002-12-06
Nirvana
·AT&T Southwest
1 edit | said by wierdo :I disagree that Lightspeed will put any pressure on cable companies, except at the very low end, which is where it does most people (and the companies) no good. "The very low end" is where most of the action is. SBC can afford "introductory" below-cost prices for a basic bundled package that should at least equal both what is currently available from the cablecos and fully meet the needs of "The very low end". Imagine what a 12 mo. $75 deal for extended cable, voice and basic DSL would mean to most people (or whatever the intro price is)
Most folks don't even know what DSL speeds they're getting. Most folks have 2-3 SDTVs. I agree that FTTN leaves very little headroom for future growth capacity though. Still, I think SBC has enough on the table here to hurt the cablecos -- at "The very low end" -- which is maybe a good 85% of the market. I'm definitely gonna try it (even though I already wish for more)
--
"Now is the time for all good men to come to the aid of their country" - and stop the NeoCons |
|
d_l
Barsoom
Premium,MVM
join:2002-12-08
Reno, NV
| Have you ever stopped to consider what the transition from ADSL to VDSL will be like? I don't think that there is supposed to be a cross talk problem between the two, but I've never seen performance graphs with the two mixed in a cable bundle.
I'd expect the VDSL speeds to suffer the most if mixed with ADSL. Will your neighborhood cable bundle have to be VDSL only to perform adequately? |
|
wierdo
join:2001-02-16
Tulsa, OK | reply to ronpin
I think you underestimate the effect of cheap HDTVs at Wal-Mart. |
|
ronpin
Imagine Reality
join:2002-12-06
Nirvana
·AT&T Southwest
| said by wierdo :I think you underestimate the effect of cheap HDTVs at Wal-Mart. Most people wouldn't know an HDTV from a TIVO -- or care. Even WalMart's $500 HDTV fails to register on most folks. They're not dumb -- just busy.
I'm feeling very charitable to suggest that SBC could be building a firm IPTV subscriber base "on the cheap" -- and then using that base to justify a final fiber FTTP network. They'd better hurry though -- 2009 is the year of HDTV. I'd think by 2012 alot of SBC's IPTV customers might start to feel cramped by the current concept of IPTV.
(again -- FTTN isn't ideal --it's just a fact)
--
"Now is the time for all good men to come to the aid of their country" - and stop the NeoCons |
|
wierdo
join:2001-02-16
Tulsa, OK
·Future Nine Corpor..
·Teliax VOIP
| If it takes until 2012 for people to be bothered by LightSpeed's lack of capability, AT&T will be very fortunate, especially since that will conincide with their normal upgrade cycle.
I still think you're underestimating the impact of HDTV, but I may have a distorted view of the market. I've heard that the cheap Wal-Mart models are selling very well. I've also read that cableco DVR placements are through the roof, and that pretty much most new HDTV boxes they're renting are DVRs.
I think we can both agree that it is the speed of adoption of HDTV that will make or break Lightspeed. If it's as fast as I think it will be, Lightspeed will be a flop of New Coke proportions. If you're correct, it will almost be a sound business decision.
And you may be right that "most people" don't know an HDTV or a Tivo from a hole in the ground, but I don't think "most people" are the main targets of a triple play deployment Verizon and SBC are in process of doing. These buildouts aren't going to be made profitable by ma and pa kettle who have 1 or 2 SDTVs and hardly would know the Internet if it bit them in the ass. They're paid for by people in big houses who have lots of nice things, soon including more than one HDTV, because God forbid they be behind the curve of stuff acquisition, even if it does put them up to their eyeballs in debt.
In some ways I hope you're right, so we can be sure AT&T will have the funds needed to fully deploy fiber in 4-6 years. In a more vindictive way, I hope they lose their shirts on it. What better way to prove to the market that half-baked B-school el-cheapo solutions will flop in the US when superior alternatives are available? (And the cablecos of today are not in any way the cable companies of 1995) |
|
