Fiber finances

When considering the many variables of deploying fiber in the loop, service providers shouldn't forget to evaluate their power architecture options

In today's local access network, there are almost as many fiber-in-the-loop architectures as providers that deploy them - and all require electrical power to perform optical-to-electrical signal conversion. The physical location of the power supply, which powers these optical-to-electrical conversions, can range in the extreme from the central office to the customer premises. But during the last 10 to 20 years, such power supplies have been migrating slowly from the CO to the customer premises, driven by the need to reduce copper loop lengths within the local loop.

Many variables affect overall costs for these fiber-in-the-loop deployments, but the focus here is on the powering and power distribution architectures, particularly given a defined optical network unit (ONU) configuration. Most studies of fiber-in-the-loop deployment costs examine variations of the size and location of ONUs. Unfortunately, because cost variables are highly dependent on the local situations and local costs, it is difficult to generalize about the costs of powering ONU configurations or architectures. For example, between rural and urban locations, trenching costs can vary from less than $1 per foot to more than $150 per foot.

For some, it might seem that power is a small contributor to the financial success of a deployment. After all, average capital costs for the powering component of fiber-fed digital loop carrier deployments are roughly 10% of the total project capital costs.

However, because operating costs are considered over a 15-year deployment or technology life cycle, the cost contribution associated with powering grows much more significant, approaching 25% of the total project life cycle costs.

As a relatively insignificant line item in the capital budget, the greater contribution of powering in construction and operating budgets can surprise many carriers. Still, given some foresight during the planning process, expenses for powering, whether initial expense, maintenance or construction costs, can be kept to a minimum.

Fiber deployments: By air or by land

In discussing the power requirements for fiber-in-the-loop applications, it is first necessary to highlight certain basic assumptions. First, for the purposes of this study, the geographic area consists of a half-mile by half-mile rectangle encompassing a mix of single-family residences, multiple dwelling units (MDUs) and businesses. Each residence and living unit in the MDU uses 1.5 lines and businesses use 4.8 lines. The costs for ONU easements are not included here because as the powering architectures change, the ONU locations are unchanged, and thus the ONU siting costs are unchanged.

However, the costs for acquiring the right of way for the remote powering system are included in this discussion, and with respect for the high variability of these costs, an average of $4000 is used. In certain high-density areas such as downtown areas, it can be necessary to locate the remote powering system in a manhole or controlled environmental vault. Access rights and installation costs in these situations can be considerably greater than the $4000 assumed in this article.

This case study examines aerial and underground deployments with a trenching cost of $50 per foot. The cabling cost, which includes the fiber and copper pairs, is assumed at $1 per foot. With aerial plant, trenching costs are replaced by pole-access costs, which are assumed at $1 per foot. Composite cables, such as those Seicor manufactures, integrate the fiber and copper pairs in a single cable, reducing cabling costs. Co-location of a remote powering system with a remote terminal allows these composite cables to carry powering and optical fiber from the co-located remote terminal and remote powering system to the subtended ONUs.

Figure 1 illustrates these costs for an aerial cable plant deployment and Figure 2 shows an underground cable deployment. The trenching costs required of an underground system make the total costs for an underground system significantly greater than for the aerial deployment. In applications where the remote powering systems are pole-mounted, the output power available often is less than that available with a ground-mount remote powering system.

The textbox on page PTN16 summarizes the methodologies of the study.

Cutting costs

Service providers should examine different powering configurations in an effort to lower the overall cost of fiber deployment. Because cabling costs are the dominant expenditures, any attempts to reduce costs must focus here first. Four remote powering systems are used in the cost analysis of Figures 1 and 2. If more remote powering systems are used, less cable is necessary. This means service providers can directly save through deploying less cable, and it reduces energy costs slightly because less cabling creates shorter remote powering system-to-ONU copper pairs on average.

Cabling still is required to connect to each ONU, so the savings in cabling are insignificant. However, if a service provider examines the cabling costs in more detail, considering splicing costs and the use of composite cable, the powering architecture can be configured to accommodate some substantial savings.

Figure 3 illustrates the estimated net present value costs for an aerial deployment, but it uses six remote powering systems compared with the four remote powering systems in the previous example. Operational costs are higher here because more remote powering system sites must be maintained. As anticipated, because the remote powering system-to-ONU distances are slightly shorter, energy costs are lower, although these savings are incremental. The total power consumed by the six remote powering systems is 3361 W compared with 3379 W in the example with four remote powering systems. Although the individual remote powering systems have a lower output power, the installation costs and equipment costs for six remote powering systems are higher than for those with four.

It is possible to extrapolate this data to a situation in which fewer remote powering systems are used in a deployment: Energy costs would increase, although only slightly, and initial equipment costs would decrease, as would maintenance expenses. With fewer remote powering systems, the power rating of the individual remote powering system would increase because the same ONU power load is supplied from fewer remote powering systems. Furthermore, fewer remote powering systems result in lower maintenance costs.

This highlights a general trend: The overall costs of powering fiber are lowered when remote powering system power supplies are used at near-maximum powering capacity. Future expansion at a remote powering system site is accommodated by the modularity of the system, which allows another remote powering system to be co-located with an existing remote powering system.

The net present value of the 15-year operational costs is most directly affected by the battery lifetime. Longer battery lifetime allows for fewer replacements of the battery string over the 15-year life cycle analysis. The strong influence of battery life expectancy on 15-year operational net present value is seen in Figure 4. This data represents the operational net present value costs for a single remote powering system.

The engine-generator alternative

As an alternative to these large quantities of batteries in the outside plant (OSP), service providers can consider deploying a standalone, curbside engine-generator, which is powered from natural gas or propane. Broadband providers are deploying versions of these engine-generators as cost-effective alternatives to batteries.

Compared with an eight-hour battery plant, life cycle cost analyses of these engine-generators show lower operating costs for an engine-generator when the power supported by the battery plant or engine-generator is greater than 2 to 3 kW. Among criteria driving growth and acceptance of curbside engine-generators in the OSP are quiet operation, agency approvals and demonstrated reliability, combined with acceptance by local governing and approval agencies.

In comparing engine-generators and battery plant, some life expectancy of the battery plant must be used, and this life expectancy affects the power rating at which the engine-generator becomes a more economical solution. Shorter battery life expectancies make the engine-generator cost-effective at about 1.5 kW while analysis with longer battery life expectancies show that engine-generators are cost-effective above 3 kW. Compared with engine-generators, standby time for batteries depends on the state of charge, temperature and discharge rate, while an engine-generator offers significantly longer standby times - up to 100 to 200 hours.The bottom line is t hat as service providers continue to roll out fiber-in-the-loop architectures, they stand to benefit from a complete examination of the architecture - including power. This examination should consider operating costs and initial costs when arriving at overall project costs. As network architectures continue to evolve, the powering issue will not disappear.