In this section we describe the development of a
network cost function from network investment data given
to us by two firms.
Firm
provided data on one of its urban service
regions containing 78,371 households.
The urban region contains 10 service areas
that each support a central office node which may
be an end office switch or a remote switching unit.
Firm
provided data on one of its rural
service regions containing 6490
households. The rural region contains 4 service areas
that also support a central office node in each area.
Although this study uses cost information from two providers, we are not trying to compare interactions between these two providers. The purpose of using the data from both firms is to develop a market cost function from a larger sample of service areas.
For both firms, the central office node can be either a central office switch or a remote access node. The remote nodes can provide single line and single party access to the first point of switching in the local exchange network. This definition of a service area is similar to the the definition that was used by Bellcore for defining Carrier Service Areas (CSA). In the Bellcore definition, the CSA can be served by Digital Loop Carriers which use T1 remote terminals to connect subscriber loops to central office equipment.[12] In our study we also include the Remote to Host cost, but the technology can be other than DLC systems.
Firm
and
Firm
have
given actual facility investment data and have
provided the investment cost data in the form of
1993 U.S. dollar replacement costs. In other words,
the cost data are the cost of purchasing new materials
at the 1993 U.S. dollar value. Other studies which
examine incremental access and usage costs have distributed
investment costs over the
average life of the investment.[5][3] They
present the data in a cost per line per month
format. Because we do not have usage costs,
we do not levelize the replacement costs of the
investments over any period. When the estimated cost function
is used in the interactive model, which is described in later
sections, the network costs are levelized over a period of 5 years.
Here we present the data in the format of total network replacement costs
in each service area.
[7]
The network elements for basic telephony access which were included in this analysis is as follows.
The details regarding how the cost data was collected is reported in another paper by Molka-Danielsen and Weiss.[7] However, we have one clarification. The Remote to Host Cost includes only the trunk costs between the a Remote Switch Concentrator (RSC) or a Remote Line Concentrator (RLC) and an end office switch. For all of the service areas that use RSC or RLC, these costs are placed under Switching Costs. The Remote to Host trunk costs for the last four service areas (K-N) could not be separated from the Distribution Link Costs, so they are included under that column.
The Total Network Replacement Cost was computed by summing the Distributed Link Cost, Remote to Host Cost, Distributed Switching and Software Cost and Drop Line Cost for each service area. These figures are presented in Table 1a and Table 1b.