Pay Now or Pay Later


History Lessons From Past Fiber Exhaust —

It’s not pretty when your organization must pay to resolve the problem of not enough fibers installed. In fact, it may be hard to believe that there was a time when fiber installation didn’t include “dark” fiber for the future. But, when we review the limitations of a young industry from an historical perspective, you’ll see why history can be unforgiving.

Learnings From the Past
In 1977, GTE installed the first operational commercial fiber optics link in North America. This set off an explosion continuing to this day. To give you an idea of the magnitude of that first link, you need to understand that it had only 6 multimode fibers operating at 6.3 Mb/s with an attenuation of 4 dB/km at 850 nm. The bandwidth of multimode fibers was approximately 160 MHz/km at 850 nm. The AT&T installation that followed transmitted at 45 Mb/s (DS-3) with spans up to 6 miles (10 km) again using multimode fiber.

The young industry had to learn how to jacket, protect, splice, and test, optical fibers. This first generation of optical fiber technology needed to progress in its sophistication over a relatively short time period. By 1983, when single-mode fiber (SMF) and laser technology became mature, the industry had to learn how to adapt from a controlled laboratory environment to handle the requirements of the outside plant.

Using standard cable jackets and cabling technology, the industry had to decide whether standard cable would work for direct buried, aerial, and, later, ducted installations. How would the cable handle moisture intrusion, high and low temperatures, vibrations, bend radius, and tension, plus armored cables? Fiber coating technology was still not mature and the 12-fiber color code (TIA-598) had not been defined yet.

In terms of purchasing, the Rural Electric Administration (REA) would allow low interest rates on fiber optic cables of up to 6 fibers. Even today this limits many of the rural telephone companies who paid the price for construction but were limited by the quantity of fibers that could be installed. The fiber “rich” cables of the early 1980s had up to 24 fibers. An additional fact pointing to the youth of the industry was that an additional fiber was added in case one was damaged during manufacturing or installation.

While early multimode fibers were cost-effective for linking central offices in the initial phases in the late 1970s, subsequent improvements in data rates using time division multiplexing (TDM) would allow migrations to 90 Mb/s and 135 Mb/s. The addition of using the lower loss and higher bandwidth 1300 nm window in the early 1980s allowed for higher speeds over longer distances using multimode fibers. At the same time cabling and splicing technologies evolved.

Also, in 1983 the industry embraced single-mode fiber technology with its tight tolerances and the new high-performance Fabry-Perot laser diodes operating at 1300 nm. Continental Telephone used ITT Telecom’s transmission equipment when they installed the first working single-mode system in upstate New York. This was later followed by MCI’s eastern corridor project which used Corning’s SMF-21 single-mode fiber with span lengths up to 26 miles (40 km). With the new SMF installed, data rates increased to 405 Mb/s and later to 560 Mb/s.

By 1987, the Synchronous Optical Network/Synchronous Digital Hierarchy (SONET/SDH) standards were completed, and worldwide communications for the TDM world were standardized by service providers up till the early 2000s. With the higher TDM transmission rates, the issue of high fiber counts was mostly a cost concern.

By 1990, fiber counts were increasing to 48, 72, 96, and sometimes the novel count of 144 fibers. (Who would ever need that many fibers?!) Most of the 144-fiber installations were for long haul or connecting central offices. As a comparison, today’s fiber counts up to 3,456 are available using ribbon fiber technology.

At that time, I recall a conversation with a local exchange carrier who wanted to build a fiber ring system with 24 fibers. I had to almost beg to get the President to agree to increase the count to 36 fibers. Even today when I visit, I still get an earful of why a larger count wasn’t installed. Sadly, they are faced with having to optically multiplex to resolve the limitations of fiber exhaust.

WDM Begins
This leads us to a third-generation fiber technology based around optical wavelength division multiplexing (WDM) technology.

In the early 1990s, WDM, using the 1550 nm window, allowed 8 wavelengths to be optically multiplexed over a single fiber. As a comparison, today up to 120 wavelengths can be optically multiplexed and transmitted. This was used originally for long haul, and especially oceanic installations, where fiber counts of 6 or 8 fibers was standard. Optical multiplexing increased the total bandwidth of the cable span. The low fiber count was due to the impact of the cost of the optical amplifiers in the spans. The good news was that the Erbium Doped Fiber Amplifier amplified all the wavelengths in the conventional band (C-band) between 1530 nm to 1565 nm simultaneously and provided a cost-effective solution.

The choices then as now were to either install another (higher count) cable on an alternate route or incorporate WDM.

Later in the new millennium, another version of WDM, called coarse WDM (CWDM), allowed for up to 18 channels. Each had 20 nanometer channels specified by the ITU-T G.694 standard. However, 5 of these channels fell into the high-water peak window centered around 1383 nm.

Today, to use the additional 5 wavelength channels, low water peak fiber per the ITU-T G.652D standard is utilized. However, unless you know when G.652D was installed, it is recommended to select channels outside the E-Band other than incur losses as high as 2 dB/km. (See Figure 1.)

Zero Water Peak Performance

Figure 1.

That said, 13 channels outside the E-band is the equivalent of 13 fibers while using the same legacy cable plant.

Solutions on the Horizon
This brings us back to the fiber exhaust issue.

FTTH fiber requirements were created from the total of how many passive optical networks (PONs) would be connected. Using the original 72-fiber-count feeder cable installed in the 1980s/1990s with 12 “dark” fibers in 1 buffer tube allowed for either 384 or 768 optical network terminals (ONTs) that could be installed.

Fortunately, the International Telecommunications Union (ITU) addressed FTTx growth through assigning new wavelengths and spectrum so that as data rates increased new wavelength assignments would coexist with legacy PONs (B-PON, EPON, G-PON). This design consideration also included the Next Generation (NG-PON2) WDM-PON where up to 8 wavelengths were assigned to address fiber exhaust and also dedicated high-speed connections.

Another bandwidth concern is the data rate required for 5G cellular transmission and the increased number of fiber drops required. 5G transmission distances and data rates will vary based on the service provider, the technology used, and the spacing requirements of the sites, which in some cases may be as short as 2,000 feet.

Six strong solutions to remedy fiber exhaust in the future are:

Solution 1: Use dark fibers as required to handle growth. (This is tricky if there is not dark fiber available given historical choices made in the early 1990s.)

Solution 2: Increase the TDM data rate to address growth.

Solution 3: Use low-cost CWDM transmission equipment with 13-18 channels, depending on the age and type of the single-mode fiber installed. These links would typically be feeder or distribution routes.

Solution 4: Transition to DWDM channels by using either dark fibers or replacing one or more of the CWDM channels.

Solution 5: Install new fiber plant.

Solution 6: Though likely costly, using an alternative physical route may provide greater flexibility and protection for the future. Designers and planners may also consider a blend of ITU-T G.652D (Low Water Peak) single-mode fiber along with a percentage of G.655 non-zero dispersion SMF (NZDS) for use with DWDM circuits, and also for leasing fibers for transport purposes.

Comparison of low-count and high-count fiber optic cable.

Comparison of low-count and high-count fiber optic cable.

Indeed, there are solutions for fiber exhaust. The method of course varies on the amount of dark fibers available, and the expected growth due to technology and consumer demand.

The best solution needs to be investigated thoroughly for both the short- and long-term. And if history has taught us anything, it is to evaluate future fiber requirements based on more criteria than the initial upfront cost.



About Author

Larry Johnson, President of FiberStory, started his career in fiber optics in 1977, and has written over 20 courses and developed 10 certifications on fiber optics through The Light Brigade which he founded in 1986. Besides his work on various standards groups, he is a member of multiple industry technology committees including the Utility Telecom Council and the Fiber Broadband Association. FiberStory is involved with the history of fiber optics, provides technology assessments to organizations including the outside plant, and represents industry organizations in fiber optic technologies. To share ideas, questions, or comments, please email or visit

Comments are closed.