CableLabs Gives Optical Fiber a Massive Efficiency Boost

You gotta hand it to CableLabs. The non-profit cable R&D organization is not about to rest on its laurels. Nor is it willing to forego a single shred of optimization possible in the broadband access network.

In its latest move, CableLabs is introducing a new optical innovation designed to make it more efficient for operators to deploy coherent optics in the access network. The work builds on existing CableLabs efforts to apply coherent optics -- a technology typically found in long-haul networks -- to last-mile broadband infrastructure. (See New CableLabs Optical Tech Promises Big Bandwidth Boost and Full Duplex, Coherent Optics Specs Advance .)

But the newest breakthrough is actually a step-function improvement in the technology that could be critical to making coherent optics viable for actual deployment.

The CableLabs innovation project is called Full Duplex Coherent Optics. Full duplex refers to the concept of bidirectional traffic over a single medium. In this case, it's not related to that other high-profile CableLabs effort, Full Duplex DOCSIS (FDX), except that both are optimizing network infrastructure by allowing traffic to flow simultaneously in two directions. With FDX, that means bidirectional traffic over a DOCSIS network. With Full Duplex Coherent Optics, that means bidirectional traffic over an optical fiber link.

Adding the full duplex part to coherent optics is important because of the way service providers are deploying fiber in their access networks. According to Steve Jia, distinguished technologist for wired technologies at CableLabs, about 20% of residential cable deployments today rely on a single fiber that uses different wavelengths for upstream and downstream traffic. But typically, coherent optical technology requires two fibers, with one laser using the same wavelength to direct signals upstream on one fiber and downstream on the other.

Further, while only 20% of cable deployments involve a single-fiber topology today, Jia projects that percentage will grow to 60% in the near future because cable operators want to use additional fibers for other purposes, like commercial services.

In other words, while cable operators plan to rely more heavily on the single-fiber approach going forward, coherent optical technology today requires a dual-fiber approach. Hence the importance of the new Full Duplex Coherent Optics solution.

CableLabs illustration of Full Duplex Coherent Optics
CableLabs illustration of Full Duplex Coherent Optics

So how does the new technology work? The CableLabs approach introduces two optical circulators on each end of an optical link to direct bidirectional traffic. Jia describes the circulator as being like a traffic roundabout for cars, but the important point is that it allows for traffic in both directions on a single fiber and it only requires one laser, rather than two. That's a massive cost-efficiency improvement.

As Curtis Knittle, vice president of wired technologies at CableLabs, explains it, these circulators will be integrated into optical transceivers in the future, making it possible for operators to deploy the technology as they're migrating to distributed access architectures (DAA).

"As an operator evolves toward distributed access architecture, they will be installing these digital optical links," says Knittle. "One option for this digital optical link between the headend/hub and the aggregation node is this Point-to-Point Coherent Optics link that we're specifying today. The technology in the specification will be integrated into an optical transceiver, a device that plugs into another device on each end. To incorporate this Full Duplex Coherent Optics, there's not going to be any change to that particular process. The transceiver potentially will have this circulator integrated and so you'll simply plug in the transceiver to the devices on each end and connect up a single optical fiber instead of two."

Get your strategic roadmap to the technology trends and business cases shaping the cable industry! Join us for the Cable Next-Gen Technologies & Strategies event in Denver on March 20-22. Don't miss this exclusive opportunity to network with and learn from industry experts -- communications service providers get in free!

Importantly, because the full duplex component has already been tested in the CableLabs' Optical Center of Excellence (OCE), there's also an opportunity now to bring the technology to market at the same time the existing CableLabs Point-to-Point Coherent Optics specification debuts. The research organization isn't willing to commit to that outcome officially, but in a blog post, Jia does note that full duplex technology "can be seamlessly incorporated into the ongoing CableLabs' P2P Coherent Optics specification effort, which will be issued in mid-2018."

That would seem to suggest that the new Full Duplex Coherent Optics work will merge with the existing CableLabs P2P Coherent Optics project. At that point, after a specification is officially introduced, CableLabs expects to move into initial trials in 2019.

There are a couple of other points worth mentioning. First, while Full Duplex Coherent Optics holds great promise for the broadband access market, it's not likely to be transferrable to the long-haul optical industry. The addition of amplifiers over long-haul links make the full duplex configuration impractical.

Second, while CableLabs is developing the new coherent optics technology for its cable members, there's no reason the technology couldn't be applied in other fiber networks by other network operators as well. In the future, this work could proliferate further with the help of telecom standards bodies like the ITU, IEEE and OIF.

— Mari Silbey, Senior Editor, Cable/Video, Light Reading

[email protected] 6/7/2018 | 6:52:14 PM
LED LED or light emitting diode is basically one kind of diode. Basically, two types of gas are used in a light emitting diode (1) Ammonia and (2) Hydrogen. So if anyone wants to know further about LED you can visit Netgear Genie support. Her you can get all recent development about the LED. LEd is hugely used throughout the world because it is user-friendly, less polluted, less operating cost.
JackZhuhong 3/12/2018 | 11:51:02 PM
FDX in coax, not fiber The FDX is in coax, which through crosstalk cancellation to send signal in both direction over the same frequency band. It needs N+0 and DAA.

In fiber, it's still seperate wavelength for up/down  in one fiber. Eg. XGS PON uplink for CMC to achieve both 10G in up/down.
comtech3 3/12/2018 | 11:32:04 AM
Re: Cablelabs You're one confused person! BW  is not the same as throughput. 2.4/1.2, is not the same as BW. Those are throughput, commonly called speed of transmission. BW is  the range of frequencies of a particular RF transport medium. Example of this is, 17- 862MHz for CATV ISPs, or 600-1400MHz  frequencies for cell phone providers. Again, what is 75/75 and 100/100 that Fios offers to its subs? aren't those symmetrical speeds, or some kind on misnomer/
majortom1981 3/12/2018 | 10:49:59 AM
Re: Cablelabs Total bandwidth on the gpon node (so to speak) is not symetrical. IF you give people above 1 gigabit it will be 2.4/1.2 Thats not symetrical.
comtech3 3/12/2018 | 10:40:21 AM
Re: Cablelabs The last time I checked, the word symmetrical means the same symetry in mathematical terms, and iso and homogenous for chemistry and biology respectively. Are we going to play pun on words? Fios has the ability to give 75/75, 100/100, or 500/500, download and upload throughput. Aren't those values symmetrical?

This simply tells me that Fios may be using the same 1550 nano wavelenth in both directions to acheive that, or some fancy modulation scheme.
majortom1981 3/12/2018 | 8:04:31 AM
Re: Cablelabs First off fios doesn't have symetrical on their ftth. They just have the speeds low enough to have symetrical.  gpon is 2.4Gbits down and 1.2Gbits up. 


Also from what I understand this is to allow symetrical down and up from the headend to the node on 1 fiber. Right now it uses 2 one fiber for down and 1 fiber for up. 


This would allow cable companies to have better use of the fiber they do have in the field. 
comtech3 3/11/2018 | 6:50:28 AM
Re: Cablelabs agreed! N+0 is achievable for cablecos given that the nodes are not too far from the actives themselves, which are spaced much closer in systems with 862MHz gears, and a wee bit further away with 750MHz line extenders.Essentally, extending fiber, and installing fiber taps shouldn't be a whole day affair that would factor in the labor cost on a hourly basis. It is the cumulative labor cost for an entire system that has over 70k subs that MAY seem cost-prohibitive. However, it is my opinion that running fiber based on the homes passed is now almost the same as running hardline coax, which possibly amounts to $500/homes passed.

If that figure is correct, or say, $750, cable MSOs are still squeezing every single digital blood out of their hardline coax,and are simply wasting money on a moribun plant (HFC) that is costly to keep on life support! Consider that they have to pay the electric companies to power their plant. Pay a lot in overtime for their maintenance technician for routine repairs, and power outages and plant breakages. With straight up fiber,the cost reduction is significant, and with the POSSIBILITY of passing on that savings to subscribers.

maurelius 3/10/2018 | 10:02:44 AM
Re: Cablelabs AT&T has shown a willingness to invest in fiber to the home for residences that are easily served by aerial. They're also doing FTTH for greenfield builds, and there's a natural amount of greenfield every year, provided we're not in a recession.

There's a significant number of customers in the Midwest who have copper in the ground - I haven't heard that AT&T is doing the expensive work of trenching/acquiring permits to serve these customers with FTTH. There's only state I know of that this has occurred in, and it's because that particular state has Google Fiber and Comcast competing with it - Georgia. You can do some Googling and find this on DSLreports.com

ATT is claiming they will pass (that can order FTTH service) 17 million homes/businesses by the end of next year - if you believe them (there's really no way to validate their claims).

The two initiatives you've identified are separate engineering projects. I can't answer your claim with respect the fiber using the same wavelength with a mixer.

Cables biggest expense in the coming years will be transitioning all the equipment on their line to n+0 - fiber to a coax splitter, _without_ any amps in the loop. You can't go symmetrical DOCSIS 3.1 without being n+0.
comtech3 3/10/2018 | 4:52:52 AM
Cablelabs Now hold on, and pardon my ignorance. Cablelabs is sending out mix signals(pun intended) here. First they're working on DOCSIS 3.1 symmetrical downstream/upstream signalling methods, while on the same token they're in an advance stage in implementing this scheme that uses the same wavelenght for both upstream/downstream with some kind of mixer. That to my knowledge, is akin to a resonant circuit.

My question is. What technique is used by say, Verizon Fios, to achieve symmetrical upstream/downstream? is Fios using two fibers with the same wavelenght to bring about that outcome? Another thing. Unless there is a massive construction boom where new greenfield homes are going to be built, I don't forsee  large investment in fiber extending beyond the backhaul. Maybe to businesses and enterprise customers, and to compliment and supplement cell towers and millimeter wave, but not residential customers.

Another thing is, ISPs will have to pump more money on top of what they have already invested in headend gears, in addition to replacing existing node circuit boards. That is costly, and entails denying customers their services while those upgrades takes place.
Sign In