This is a picture of the Ben Johnson Associates logo.
Welcome Firm Clients Experience Expertise Personnel Services

Site Home

Services Home


Work Product

Telecom Economic
Cost Model

Telecom Economic Cost Model

Using GIS in the Telecom Model

In preparing cost estimates for a set of wire centers, we typically use an entirely data-driven, largely computerized approach. We begin with telephone numbers and addresses from the white page listings, distinguishing residence and non-residence (business) listings. Since detailed white page listing data will be available for all wire centers studied, a high degree of consistency is ensured. To the extent feasible, we identify the exact geographic location (latitude and longitude) of each listing, supplemented with address matching using zip+4 and street segment data where possible. (The overall accuracy and completeness of this process is more than adequate for most purposes; however, for allocating universal service funds, further improvements could be achieved, particularly for remote customer locations in rural areas. Where a very high degree of accuracy is needed, the rural GIS data can be improved by digitizing satellite photos and/or gathering precise customer location data on site, using hand-held GPS equipment.)

This is a picture of the Telecom Cost Model logo.

Once the listing data are geo-coded, we match them to the approximate boundaries of the wire center serving areas (from ExchangeInfo) to define the geographic boundaries of each distribution area, to locate the latitude and longitude of each DA node, and to connect each DA node to its respective wire center through a series of feeder segments. This process, entirely data driven is accomplished using ArcInfo GIS software on a largely automated basis (ArcInfo software, developed by Environmental Systems Research Institute of Redlands, California is the industry-leading GIS software). The results can be shown on maps of the feeder segments and distribution areas, with end-user sites pin-pointed.

The schematic feeder network designs can be entirely derived from the listing data for each wire center, without any reference to the existing network configuration other than the wire center location. However, the Telecom Model is not limited to this pure "scorched node" approach but can also incorporate various aspects of the existing network, if desired. For example, the GIS data can reflect the locations of the serving area interfaces between the feeder and distribution cable within the existing network. In fact, the Telecom Model provides almost complete flexibility with regard to the geographic aspects of the network.

The GIS approach used in the Telecom Model provides additional benefits in developing accurate cost estimates which do not rely upon embedded cost data. For example, we can use the ArcInfo GIS software to organize and summarize relevant data concerning not only spatial characteristics of the network, such as feeder segment length and distribution area size (square miles), but also indicators of the soil texture, bedrock depth and hardness, groundwater depth, and road density along each feeder segment route, and within each distribution area. This information is used in the Telecom Model to more precisely estimate the cost and the type of structures (aerial/underground/buried) of the feeder and distribution cables within each portion of each wire center. As a result, the Telecom Model can develop cost estimates which are precisely matched to local conditions, without the need to rely upon historic cost data.

We must emphasize, in this regard, that the GIS data are inputs to the model, not outputs from the model. The Telecom Model can use virtually any geographic data source, provided it can be converted into a GIS compatible format. As better data sources are developed (e.g., exact geographic locations of customers in remote areas) the accuracy of the Telecom Model cost estimates can be readily improved without further changes to the model itself.

The Telecom Model can also use the incumbent LEC's actual route data. In that case, it is reasonable to ask how our results would be different from the LEC's embedded cost data.

Assuming that some or all existing LEC facility locations are used in the model, there will still be numerous differences between the resulting cost estimates and the embedded costs--cable sizes, cable technology (copper/fiber mix), as well as location and type of remote electronics.

Cable Technology Choices

A long-run cost study should consider three distinct technologies: analog copper, Digital Line Carrier (DLC) on copper and DLC on fiber. In this regard, the Telecom Model is far more flexible and powerful than competing models. The user can choose between analog copper and DLC fiber using either of two different engineering criteria (based on total loop length or based upon cumulative feeder length). However, the user can also choose an all-copper network, using a combination of analog and DLC technology, or can choose to model a network which uses fiber in all of the feeder segments. Finally, the Telecom offers a cost minimization option. With this option, the model selects and deploys the cost minimizing technology, or combination of technologies, on each feeder segment, and it identifies the best locations (nodes) for remote DLC electronics. It is worth noting that the least-cost technology selected by the model may differ, depending upon the share of the market served by the carrier, as well as other factors. The model identifies the least-cost option by checking every possible combination of technologies, segment by segment and node by node, until it identifies the the lowest cost, most efficient solution. The approach used is similar to the manner in which "Deep Blue" solves a chess problem by rapidly and repetitively considering all the possibilities.

Spare Capacity or "Utilization Factors"

Some models greatly overstate the cost of spare capacity, because they don't actually model the additional costs associated with utilization factors below 100 percent. Instead, they calculate the weighted basic network element (BNE) cost assuming 100 percent utilization, and then divide this figure by its assumed utilization factor. This procedure ignores the fact that economies of scale or density can apply to specific facilities, causing the cost to increase less than proportionally as the size of the facility increases to provide for spare capacity.

In the Telecom Model, utilization factors are used to separately increase the size of each facility or piece of equipment. In turn, the model estimates the cost of this larger item, incorporating the necessary spare capacity. This results in much more precise cost estimates, which are not biased upward.

To size facilities, the model uses separate utilization factors for fiber electronics, switching equipment, various categories of cable, as well as other categories. Use of these factors ensures that an adequate amount of spare capacity is available for each type of facility. A long-run study would normally use relatively high (idealized, or optimum) utilization rates, consistent with the underlying concept of a long-run planning horizon, in which factory size is adjusted to accommodate the anticipated volume of production. It should be noted that the actual amount of spare capacity and the effective fill factor (reflecting the actual ratio of working facilities as a percentage of total installed capacity) are developed within the model where the model takes into account the lumpiness of specific facilities (e.g., cable). The effect of this modeling process is generally to increase the amount of spare capacity (and reduce the effective fill factor.)

Please follow the "Contact Us" link below to get information on how to call, email, or fax us your questions about the Telecom Model and how to get it working for you.

| Sitemap | | Job Opportunities | | Contact Us | | Privacy | Link to

Copyright ©1996-2002 Ben Johnson Associates, Inc.® All Rights Reserved