Fiber-to-the-home’s status as the pre-eminent bandwidth delivery medium, coupled with the emergence of IP video, is driving higher bandwidth requirements across America, including into rural locations. A seldom-mentioned byproduct of these rural deployments is the need to link remote central office and hub locations over substantial distances. The capacity of those links will define the provider’s ability to deliver next-generation services in the last mile, and the distances those links support may introduce issues which were previously the exclusive domain of long haul backbone providers. Here, we will examines those issues and define the points where they come into play for a rural FTTH deployment.

There are a number of relevant factors that contribute to the success of any FTTH deployment. For the optical media itself, however, the most significant parameter between the central office and the home is attenuation. Attenuation is commonly understood by last-mile engineers, and is reasonably easy to both calculate and measure. However, a long link to a remote central office or hub site has the potential to introduce another consideration: dispersion. The importance of dispersion is determined by a combination of both the distances involved and the amount of bandwidth being transported. Figure 1 illustrates where the limiting parameters begin to shift.

Dispersion: The Good, the Bad, and the Ugly
To define dispersion, let’s look at optical transmission at its basic level. Not all wavelengths of light travel through an optical fiber at the same speed. Like a plume of smoke from a stack that gets broader and harder to recognize as it travels farther from the stack, an optical pulse broadens as it travels down the fiber. This can result in detection errors at the receiver.

In single-mode fiber, most of the broadening is due to a property called chromatic dispersion. At 1310 nm, single-mode fiber has very little chromatic dispersion. However, at the wavelengths used for longer distance transmission, such as 1550 nm, the dispersion can be significant when distances are over 60 kilometers and either analog video is being transmitted or data rates are greater than or equal to 10 Gb/s.

There are three options for addressing dispersion concerns:

Option 1
Improvement in laser transceiver technologies. Improvements in laser technology are certainly a possibility and various vendors indicate that new solutions are on the near horizon. However, the price and availability of new transceivers are unknown variables. Therefore, this analysis is limited to the known trade-offs between dispersion compensating modules and specialized fibers.

Option 2
Retrofit solutions for conventional single-mode fiber, in the form of dispersion compensating fiber (DCF) modules. A DCF module contains a specially designed optical fiber that reverses the chromatic dispersion that develops as a signal propagates down an optical fiber. These devices typically have an insertion loss of about 5-10 dB when removing the dispersion generated in about 50-100 kilometers of conventional single-mode fiber. Although relatively costly, DCF modules are convenient, as they can be added after the optical cable has been deployed.

Option 3
Use of optical fiber with modified dispersion characteristics. The logical current alternative to the DCF module is to deploy non-zero dispersion fiber (NZDF). Using NZDF allows transmission of one hundred fifty miles at 10 Gb/s with minimal system impediments due to chromatic dispersion.

When an application calls for some level of dispersion mitigation, NZDF has two advantages. First, NZDF is less expensive than conventional fiber retrofitted with dispersion compensating modules. Second, attenuation of the optical link with NZDF is less than the attenuation of a conventional fiber with the DCF module. The second advantage also has a financial bearing since it greatly reduces the need for amplification.

Rural FTTH Deployment Realities
The most immediately relevant dispersion-affected application for the rural independent communications provider is the crossover point between 10 Gb/s and 60 kilometers. A distance of 60 kilometers between central offices or remote hubs is not uncommon for a rural carrier. And a 10-gigabit backhaul upgrade is a rational expectation during the first few years of a FTTH deployment.

As an example, Blue Ridge Mountain EMC is an electric cooperative deploying Internet and IP video in rural North Georgia and Western North Carolina. Their architecture includes node sites where a router receives a backhaul connection and provisions the equipment that transports services via FTTH to subscribers. Those node sites are occasionally in excess of 60 kilometers from the primary equipment room, and Blue Ridge Mountain EMC needed to evaluate 10 gigabit upgrade options to the nodes before the first year had elapsed on their deployment.

The reason for the rapid upgrade to 10 gigabits stems from the current demands Blue Ridge EMC required for support of IP video. This cooperative is not unlike other rural providers. They must deal with the ever-increasing level of bandwidth usage among subscribers, and the low number of fibers used to support the remote routers.

Interestingly, historical trend lines indicate a growth in nominal data rate usage among Internet subscribers at nearly 42 percent annually over the past several years. This growth rate may plateau in rural areas where there is relatively little competition among providers. Nevertheless, 10-gigabit backhaul requirements can materialize even under much slower bandwidth growth conditions, as when IP video drives a beginning nominal data rate of 30 megabits-per-second for new subscribers.

In this case, Blue Ridge EMC had 4 fibers providing backhaul to a remote router at a node site. If the node site supports 1,000 subscribers at a nominal data rate of 30 megabits-per-second and at least 90 percent of all traffic routes through the backhaul fibers, then Blue Ridge EMC must begin with a need to support just under 7 Gb/s for each backhaul fiber during the first year. If a rural provider like Blue Ridge EMC started under those circumstances and actually realized the historic 42 percent annual growth in the nominal data rate, the increase in capacity requirements for the backhaul would resemble the bar graph in Figure 2.

It may be unlikely that we will see serialized transmission (a single wavelength on a single fiber) of 150 Gb/s in the near future. Most experts agree that such higher-level data rates will be marked by a transition to CWDM or DWDM technologies.

Nevertheless, outside plant infrastructure is deployed with the intention of supporting operations for as long as 40 or 50 years. Certainly, therefore, we can expect 10 Gb/s backhaul links to be ubiquitous in FTTH, while 40 Gb/s serialized transmission may be common before a backhaul fiber installed today exhausts its utility.

Finding the Most Cost-Effective Approach
While a 10 Gb/s link requires consideration for dispersion at 60 kilometers, a 40 Gb/s serial transmission link is challenging. For example, after about 2 miles of transport on conventional single-mode fiber, the transmission is impaired due to chromatic dispersion. It is entirely possible that 40-gigabit requirements will be met through CWDM or new laser technologies that circumvent the dispersion issue. Nevertheless, an awareness of dispersion will still be necessary for those charged with selecting the next upgrade technology. Therefore, the cost trade-offs in the currently available technologies of NZDF fiber and DCF modules must be weighed into the analysis of any new capital investment in backhaul fiber links.

In most instances where dispersion is an issue from the beginning of a deployment, NZDF is a more cost-effective option than DCF modules and possible amplification. However, if the bandwidth requirements and distances make dispersion mitigation just a possibility on the horizon, DCF modules have the advantage in that they can be installed as needed. Thus, the DCF module approach allows for deferred expenditures until the day when the investment is absolutely needed.

The Impact on Signal Loss and the Bottom Line
In most circumstances, the number of fibers used to support the remotely located routers in an FTTH deployment will be very low - probably no more than 4 or 6. Furthermore, it's likely that the cable that houses those fibers will be required to perform the additional duties of provisioning optical splitters or supporting drop locations for customers along the route. For those applications, a certain number of traditional G.652D fibers are not only desirable but also mandatory.

Thus, all of the costs associated with dispersion in the FTTH backhaul scenario are significantly lower than the costs reflected in a traditional long-haul deployment. After all, if we need NZDF fibers, we will need only a few of them. And, likewise, if dispersion compensation is required, it will be for only a very few fibers.

While the cost trade-offs and time value analysis of NZDF versus DCF modules is critically important, the issue of attenuation cannot be forgotten. As an example, Figure 3 compares two 50-mile optical links. One has conventional single-mode fiber with a DCF module and the second uses NZDF. Figure 3 clearly shows that the total loss for the NZDF link is 4 dB less. In many cases, this can result in the deployment of fewer and less powerful optical amplifiers in the system, which greatly reduces capital costs, operating costs, and engineering complexity. This impact of DCF module deployments on attenuation helps bolster the case for NZDF where dispersion comes into play.

When backhaul bandwidth growth, a migration into rural applications, and the relative uncertainty of cost surrounding new high-bandwidth transmission technologies are considered, the case for a few NZDF fibers in a backhaul cable begins to look attractive. The typical backhaul cable being used for FTTH is a multipurpose component, and it is likely that less than 12 fibers in any cable would need to be NZDF in order to support a dispersion-impacted application.

Overall, that relatively low number of fibers has a negligible impact on the total cost of network construction, but can possibly mitigate a tremendous number of potential headaches. Rural FTTH carriers should give that option serious consideration whenever links reach the distances and bandwidths where dispersion could have an affect.
 

About the Author
David Mazzarese has more than 15 years experience in optical fiber manufacturing, R&D and optical systems engineering. He is technical manager of optical fiber systems engineering supporting the entire optical path for OFS and can be reached via email at dmazzarese@ofsoptics.com. 

About the Author
Guy Swindell has more than 10 years experience in field engineering, applications engineering, and installation project management. He is manager of field applications for OFS and can be reached via email at gswindell@ofsoptics.com.

What’s your take on this subject?  Leave a comment and get the conversation going.