Getting More Out of Telecom Fiber

HANK HOGAN, CONTRIBUTING EDITOR, hank.hogan@photonics.com

The rising amount of data generated by people and machines around the world travels largely by fiber, with some of it moving thousands of miles in a hop. For some long jumps, new ultralow-loss, large effective area fiber is going into the ground and under the sea, leading to higher performance. There also are new signaling modulations and other innovations, all aimed at boosting bandwidth and cutting the cost of moving data.

But upping transmission power, which leads to a longer reach, or faster speeds in other situations, is not an option when it comes to fiber, according to Sergejs Makovejs, market and technology development manager at Corning Inc. of Corning, N.Y.

For every type of fiber, there is a power level that gives the best results.

“If you keep increasing the light power further, your performance will go down,” Makovejs said.

This is because fiber is a nonlinear medium, which means brighter lasers can lead to crosstalk between channels and other performance degrading effects. Thus, there is a fiber-dependent optimum light power point, a level at which performance is maximized. Transceivers and other system components are typically set at this point, Makovejs said.

What does work to increase bandwidth and cut costs is ultralow-loss, large effective area fiber, which has long been used in submarine applications, according to Makovejs. Such fibers have special manufacturing techniques and materials such as a pure silica core instead of a germania-doped one. These improvements drop attenuation from 0.2 dB/km to less than 0.17 dB/km, which translates to a 3-dB gain in system performance per 100 km of fiber. The 0.17 dB/km attenuation level is the threshold for the ultralow-loss fiber designation, per industry convention.

Such improvements are important when data rates rise. Corning calculations show that going from 100-Gbps transmission rates, the current state-of-the-art, to the soon-to-be standard 200 Gbps can drop the signal reach from a few thousand kilometers to several hundred for a fiber with typical loss characteristics. With ultralow-loss, large effective area fiber, the point at which signal regeneration is needed can be significantly extended.

That is beneficial because there is a trend for signal regeneration sites to be eliminated. One example of this can be found in submarine cables, which in the past terminated on the shore with the data then moved to a data center. Thus, there were two regeneration points: one on the shore and another in the data center. Now, new submarine cables are more often terminating in a data center itself, thereby cutting the regeneration points in half and producing cost savings.

“In an optical fiber, you typically have approximately 100 wavelength division multiplexed channels. With regeneration, you would have to regenerate every single wavelength, and that becomes very, very expensive,” Makovejs said.

Spools of optical fiber. Increasing data demands are leading to new telecom fiber going into the ground and under the sea. Courtesy of Corning Inc.

He added that in addition to different materials and specialized manufacturing techniques, the light-carrying core of ultralow-loss, large effective area fiber is about 125 µm². That is larger than that of standard single-mode fiber, which has a cross section of 82 to 83 µm². Single-mode fiber is used in long-haul applications because it offers greater reach. A larger light-carrying area means that more power can be sent through the fiber before nonlinear effects cause problems, and so a larger cross section thereby increases reach, data rate or a combination of both.

5G networks driving data traffic

More bandwidth and lower cost per transmitted bit are needed to deal with the rising data traffic projected to come from fifth-generation (5G) mobile services. The associated mobile networks provide download speeds of up to 1 Gbps, 200 times the currently available through-put. The latency, or lag time, of data delivery will be in milliseconds. The result of this increased performance is that data consumption will rise, with more video being sent to mobile devices. ABI Research forecasts that worldwide 5G service revenues will reach $247 billion annually by 2025.

The emergence of smart cities is a major driver in increasing network bandwidth.

These 5G networks are in the process of being deployed, with the net result being that the demand for telecom traffic is expected to increase. That and other applications are why the global fiber optic market is expected to grow at 9.8 percent a year, rising from $3.13 billion in 2016 to $5 billion by 2021, according to Abhinav Pachauri, an analyst with MarketsandMarkets Research Pvt. Ltd., a business-to-business market research and consulting firm.

The global fiber optics market was estimated at $3.1 billion in 2016, with market growth attributed to increasing compliance on industry standards. Growth of application areas offers lucrative opportunities for the market players through 2021, resulting in a nearly 10 percent per year growth rate worldwide. Courtesy of MarketsandMarkets Research Pvt. Ltd.

“Telecom accounts for approximately 50 percent of the fiber optics market share in terms of volume throughout the forecast period,” Pachauri said.

Much of this demand is due to rapid growth in developing counties in the Asia-Pacific region. In this area there may be brand-new runs, which means that ultralow-loss, large effective area fiber could be used in terrestrial settings. Given that fiber going into the ground is expected to last 25 years, extra capacity and higher-performance fiber may be installed to accommodate future growth.

Another source of increasing telecom demand comes from data centers. These are major sources and sinks for traffic that travels over fiber, sometimes to and from destinations within the same region and sometimes to and from locations around the world. Service providers and companies that have their own mega data centers — such as Facebook or Google — want faster speeds, said Jonathan Jew, president of cabling infrastructure specialist J&M Consultants Inc. of San Francisco. The company’s work is within data centers, with some of this involving connectivity to telecom service providers and carriers.

Jew noted that the pace of development of faster speeds is dictated by advances in standards, particularly IEEE 802.3, which, among other things, specify transceiver data rates. In addition to speed, he said, the cost of a fiber-based solution is another critically important parameter. He added that increasing shipment volumes of faster devices will hopefully result in overall lower costs.

He also noted that having telecom fiber already in place, as is the case in the developed world, is not a bad thing.

“The available fiber under the ground is single-mode. I don’t see that we need to replace it, though we may need to increase fiber counts and pull new fiber as needs arise,” Jew said. “Having the fiber under the ground is a good thing as it speeds deployment.”

If new fiber is needed, however, putting it in the ground in the United States, Europe and elsewhere in the developed world presents a problem, according to Ron Johnson, senior director of product management at Cisco Systems Inc. In such locations, the process of getting a permit for new cabling is difficult, so most of the time carriers instead achieve higher speeds and lower costs by changing transceivers and other optoelectronic components.

For example, the amount of data carried in the optical stream can be increased by using a more complex encoding scheme. However, going to a more complex scheme places a greater burden on the electronics and photonics. Devices must translate an incoming bit stream into a series of modulations such as shifting the phase of a reference signal, transmit it, receive it and then translate it back into bits to send on to the rest of the network. As speeds increase and modulations become more complex, the components must do this entire process faster and faster.

Faster processing

Fortunately, semiconductor advances mean that processing and signal manipulation become more powerful and affordable over time, a trend that Cisco takes advantage of, Johnson said. For instance, faster processing makes it possible to mix modulation schemes to tailor reach and data rates for a given application.

Johnson said that a 100-Gbps data rate might have a range of 10,000 km. A 600-Gbps data rate, on the other hand, might only have a reach of a couple hundred kilometers. Before recent digital signal processing improvements, telecom carriers had to pick one modulation scheme to use. Now, this balancing of reach against rate can be made and changed in a way that allows more data to travel down a given fiber for less money.

“That trade-off is giving carriers the ability to leverage the distance they have to go in the network to get more bandwidth in the fiber. You can do that on a wavelength by wavelength basis,” Johnson said.

Semiconductor advances have cut the power used by digital signal processing from 100 W per channel down to 20, thereby lowering power consumption significantly and reducing operating cost. There also have been expansions of the wavelengths used, with the band from 1565 to 1625 nm joining the standard 1550 nm long used in telecom. The addition of extra wavelengths can effectively double fiber capacity, according to Johnson.

Such innovations outside the fiber itself can go a long way to meeting the needs in telecommunications for more bandwidth. That can be combined with ultralow-loss fiber, fewer regeneration points and more to achieve the industry’s goal of cutting costs.

“The cost per bit is very much tied to the bandwidth you can get per wavelength,” Johnson said, “and how many wavelengths you need to do the job.”