Verizon Spells Out 100 Mbit/s

Verizon Communications Inc. (NYSE: VZ) is admitting that its 100-Mbit/s plans will come with a few caveats.

That's not to say 100 Mbit/s, even a bit diluted, is child's play. Verizon has come forward as the only major U.S. carrier to say that delivering 100 Mbit/s to the individual home is on its radar. (See Verizon Leads the Great 100-Mbit/s Bandwidth Race.) It's just that getting there won't be easy.

Luckily, Verizon won't have to saturate its network with 100-Mbit/s lines. "At any stage we don't have 32 people looking for 100-Mbit/s service," says Vincent O'Byrne, director of access technologies at Verizon. "We would do some oversubscription."

That means Verizon is betting that whenever 100 Mbit/s first becomes a reality, the demand will be minimal. "It's not likely that everyone would want it," says O'Byrne. "It would be a small percentage of people in any neighborhood."

Good thing, too. Verizon says that most of its central offices serve about 15,000 to 20,000 homes. Simple math would dictate that to bring 100 Mbit/s to each of those homes would require at least 1.5 Tbit/s of bandwidth going into each central office. That of course is not how things work now.

While Verizon's migration from BPON to GPON will play a big role in the movement towards 100 Mbit/s, traffic management will be just as significant.

"When you oversubscribe, you have aggregators behind the OLT [optical line terminal] which take advantage of the dead space of the transmission from one OLT to another," says O'Byrne. "It is taking advantage of statistical multiplexing and the fact that not everyone is accessing that 100 Mbit/s at the same time."

In the long run, Verizon is considering the possibility of upgrading its network with Wavelength Division Multiplexing (WDM), or WDM-PON. This technology sends a full wavelength to each house, greatly increasing the potential capacity of each fiber.

"We're making sure that when we deploy our GPONs, that we have the capability when there is a need, to increase our network's output. Part of our overall strategy is to make sure our network can evolve to WDM," says O'Byrne.

For now though, O'Byrne says that Verizon is expecting GPON to have a long lifespan that won't get cut short by WDM-PON.

"There are different substantiations of WDM," O'Byrne says. "There are some variants that allow you to evolve off the present GPON, and they sort of become another GPON channel."

But even with network capacity upgrades, oversubscription, and multiplexing, signing up for 100-Mbit/s service won't mean you'll get that bandwidth all the time. "100 Mbit/s isn't guaranteed," O'Byrne says. "If you ran a test on the network, it would read 100 Mbit/s, but if enough people were demanding it at the same time, we couldn't guarantee each would get the 100 Mbit/s."

— Raymond McConville, Reporter, Light Reading

Page 1 / 3   >   >>
bollocks187 12/5/2012 | 2:57:21 PM
re: Verizon Spells Out 100 Mbit/s Surewest in Sacra. CA deploys Active Ethernet. Also several carriers in EMEA.
jepovic 12/5/2012 | 2:57:24 PM
re: Verizon Spells Out 100 Mbit/s Yes, I have been at the US east coast.

Are you so naive as to think that a company like Verizon base their decisions primarily on rational network cost calculations?

Of course they think very seriously about the regulatory aspects as well, which was one of my points.

The other point is that they may have been fooled by the vendors. This happens all the time! A couple of high-level meetings on the golf course can "make" the decision about technology and vendor from the top. Alternatively, efficient marketing towards the right technical "gurus" can "make" the decision from the bottom.

After all, these are terribly difficult questions with an endless amount of parameters to estimate. In the end, gut feelings will be more important than numbers. As has been shown in endless organizational studies, business cases and number crunching is usually done as a justification after the actual decision (but before the formal decision).
rjs 12/5/2012 | 2:57:25 PM
re: Verizon Spells Out 100 Mbit/s As JEPOVIC wrote in post 19

" ....

I don't understand why this model isn't used by eg Verizon on the east coast.

In suburbs, this model is more tricky. Still, PON seems awfully complicated, poorly standardised and anything but future-proof. I still think it's a dead end. "

I believe the above was a rhetorical question, but if not my opinion is as below.

I believe the answer is that everybody will then know that technology is not the issue with 100Mbps deployment. PON gives a good barrier to entry for
monopolistic walled gardens.

Also, throughput and toggle rate should not be confused. Throughput's upper limit is defined by the toggle rate. With a 100Mbps (toggle rate) p2p ethernet, the throughput can be much lower due to statistical muxing, and once the back bone is ready to handle a throughput of 100Mbps per subscriber upgrades are easy.

Technology is not the bottleneck, FCC and monopolies are the real bottleneck.
Remember, if it were not for VoIP we would still be filling the coffers of ILECs for 10c/minute calls. ILECs will not change unless their revenue is threatened.

Duh! 12/5/2012 | 2:57:26 PM
re: Verizon Spells Out 100 Mbit/s Quote:
I don't understand why this model isn't used by eg Verizon on the east coast.

In suburbs, this model is more tricky.
End Quote:

Have you ever been to the US east coast? Do you actually think that a substantial portion of the population in Verizon's territory lives in MDUs?

More to the point, are you so naive as to think that Verizon hasn't modelled, analyzed, confidence tested and analyzed again their opex/capex model for all kinds of architectures?

What is it about this topic that brings out all kinds of people with opinions but no facts?
tailpin 12/5/2012 | 2:57:27 PM
re: Verizon Spells Out 100 Mbit/s I'm trying to figure out these next-gen PON abreviations. The most common seem to be PON and lol. What's a lol-PON?
jepovic 12/5/2012 | 2:57:27 PM
re: Verizon Spells Out 100 Mbit/s "* Active Ethernet does not scale for mass-market deployments in moderate-to-low density areas. Too many fibers coming in to the CO, too many fibers to fit into existing duct, too many fibers to repair quickly in event of a dig-up. And if you go to remote nodes in the field to eliminate those problems, you're back to the powering and maintenance issues that come along with VDSL and ADSL2, but without the cost avoidance of reusing copper distribution and drop."

You're talking about P2P fiber from the CO, which indeed is pretty rare except in greenfield deployments. I'm referring to fiber to the basement (FTTB). In this case, the amount of fiber is similar to PON. Yes, you will have distributed active equipment, but it is cheaper and way more powerful than the distributed PON equipment. Power is regular 220 AC, cooling is done without fans. All that is needed is a room in the basement.

There are other advantages of using Ethernet switches as well: Efficient multiplexing on packet level, simple well-known technology (easy to train support staff, field technicians), tons of vendors, simple physical interfaces etc.

I don't understand why this model isn't used by eg Verizon on the east coast.

In suburbs, this model is more tricky. Still, PON seems awfully complicated, poorly standardised and anything but future-proof. I still think it's a dead end.
bollocks187 12/5/2012 | 2:57:27 PM
re: Verizon Spells Out 100 Mbit/s Answer - None !

It's all a 'media' show. Verizon does not know what it is trying to do - it is all reactionary.

BPON(ATM) then GPON..... seed the future with WDM-PON.......then BS-PON - lol

bollocks187 12/5/2012 | 2:57:28 PM
re: Verizon Spells Out 100 Mbit/s 7,

What vendor makes a profit with the GPON solution ?
redface 12/5/2012 | 2:57:28 PM
re: Verizon Spells Out 100 Mbit/s Actually the WDM-PON laser is just a coolerless low cost FP laser with a 0.1% AR coating on the front facet. Reflective SOA is just another name for the same thing, more or less. They are both based on laser injection locking, with a seed wavelength signal provided by a wavelength sliced broadband light source. Wavelength slicing is done using AWG. A batch AR facet cost does not increase manufacturing cost very significantly. Otherwise it is identical to a diplexer in optical path and build. The cost of manufacturing this ONT in the high volume manufacturing limit is about the same as that in diplexer, and cheaper than a triplexer which is frequently used in TDM-PON by Verizon. Of course the AWGs used in WDM-PON will increase the WDM-PON component cost significantly. However, the per channel AWG cost is getting to $30-40 these days so it is not a show-stopper either.

paolo.franzoi 12/5/2012 | 2:57:28 PM
re: Verizon Spells Out 100 Mbit/s
As said, TDM PON can be 2 wavelengths AND can be done with an FP to match the distance. If you want to do B+ distances with Novera, you are not using an FP laser.

The way to get to FP with GPON is to use a bit more aggressive rules on the ODN design AND use FEC. You will find you can use FP lasers in the ONT and meet the distances required. Of course, Verizon won't allow that small a margin in their design.

And the $30 cost for the AWG is about 1/3 the Price of an indoor diplexer ONT as shown that Novera makes in either GPON or EPON. So, yes it is significant.

Page 1 / 3   >   >>
Sign In