On Distributed Communications Series
VI. Mini-Cost Microwave
VIII. Cost Estimate
The following cost breakdowns are included to indicate the approximate system price range. These costs include sufficient profit for each operation so that a commercial organization would find all business worthwhile.
The figures are based specifically upon the mass production approaches suggested and are for the specific system discussed.
Care should be taken in comparing these numbers with those of non-commensurate systems. (The temptation is severe when it is realized that the resulting price estimate for the initial installation is less than the equivalent of four months' rental of the cheapest commercial alternative--and one which has a lower data rate.)
Cost of Antenna Siting and Erection
Siting and erection cost estimates are based upon the performance of those tasks described in detail in Table II. Most likely, the contractor will be a small organization, in which a high level of work performance is demanded. Pay would be commensurate. The speed of emplacement of the towers should serve as a good metric for basing incentive payment bonuses. The performance costs breakdown is shown in Table III. The high burdened-salaries and the heavy overhead-loading may appear to be excessive for the relatively simple and routine job to be performed, but profit and management expenses are included in the overhead-loading. The total workdays' figures are for "usable days"--assuming that laborers are paid whether the weather permits work or not. Training costs are not included since all tasks can be learned on the job.
The costs are presented on the basis of dollars per function per working site. The total cost is about $1317 per site, divided approximately equally between the costs of selecting the site and those of erecting the antenna and the electronics.
Cost Breakdown of Transmitter/Receiver Package
The costs for the plug-in transmitter/receiver unit are broken down in Table IV, amounting to some $1153.20 per unit (with two required at each site).
Total Cost for the Complete Link
Table V lists the costs of the complete repeater station and the total cost on a per-mile basis.
This price estimate will be increased by throwing in a blanket contingency safety factor, thereby arbitrarily raising the total price to around $400 per mile; this figure will be used when computing total system cost in ODC-X.
Because we do not have a good way of predicting the mean time between failures (MTBF) for the all-solid-state system, we shall use a pessimistic upper-bound estimate and work backward.
A minimum figure of 100 days MTBF per relay station has been tentatively assumed (in order to maintain desired network performance with a 24-hour average time of repair). Thus, a maximum average of 3.6 service calls per year per station may be expected. Because of the amount of time spent traveling to the site for service calls, service might be best contracted out to local TV servicemen. Immediate repair is not mandatory. The cost per service call could be as high as $100; the total cost per year, then, would be only $18 per mile. If in-house servicemen are used, the number of such technicians needed does not fully hinge upon the reliability of the equipment. For example, if the stations exhibited almost perfect reliability, the few servicemen absolutely required would be so widely spaced that they would be spending almost all their time traveling. It would seem better to have the same technicians who service the Switching Nodes and Multiplexing Stations also service the links, in order to raise the density of the servicemen and decrease travel time. However, even if the average repair requires an unhurried 24 hours, the system would still be highly reliable.
Consider Table VI which shows the availability factor of at least one link for a Switching Node having up to four links (six or more links is closer to the value that may be anticipated in the actual network). Two values of link reliability are assumed in Table VI, 0.9 and 0.95. It can be seen that extremely reliable network access does not really require highly reliable links, and we can get by with delayed maintenance.
 This seems to be quite conservative for a small, rather reliable, all-solid-state unit having about 40 transistors and two sources of power.