ECONOMICS OF TELEPHONE RELAY AlM'LirATIONS 219 



surprising that relay use has an enormous influence on the cost of the 

 central office — not just because of relay purchase cost but also the 

 rela3''s influence on other office factors. For example, the size of the 

 power plant and the total number of eciuipmcnls for common control, 

 which (k^pend on the functioning time of tiie relays, are decided by the 

 rela}' characteristics. In the application of relays, then, it may well 

 prove that the largest economies can be realized by spending a little 

 more for each relay in the beginning in order to save still more in the 

 cost of the power plant, common control equipment, size of the hnildiiig, 

 and so forth. It has been found that attention to ways of optimizing flu; 

 application costs for relays can lead to an appreciably lower cost central 

 office, and this paper will illustrate a few cases of how such a problem is 

 approached. Because the telephone system is so large, the influence of 

 each relay, taken in total, is also large. Fortunately, at Bell Telephone 

 Laboratories, it has been possible to take the over-all view of the sul)ject 

 through familiarity with all phases of the relay application probUnii; 

 and large economies which will eventually benefit the cust(jmer are 

 resulting. 



The basic problem to be discussed is how best to realize maximum 

 economy of the central office so far as relays and their uses are concerned. 

 As in most engineering problems, it is necessary to evaluate and com- 

 promise between oppositely varying cost effects of such things as effi- 

 ciency, manufacturing cost, amount of equipment, ease of mounting 

 and wiring, maintenance in all its aspects, and the like. The end result 

 of each of these variables is its economic effect on the office as a whole, 

 and they can onl}^ be compared if their values can be stated on a com- 

 parable basis. It has been found that each such effect may be considered 

 as an incremental cost over and above a reference cost for its particular 

 ideal condition. If properly chosen so as to be independent, then all such 

 incremental costs may be added. They then represent the net "cost 

 penalty," compared to the design with all ideal conditions taken to- 

 gether, and describe the merit of the design. They can also be used to 

 find the optimum design. The methods apply generally to many other 

 similar problems. 



BREAKDOWX OF THE PROBLEM 



Consider first how relays are applied in the telephone switching sys- 

 tem. To the greatest extent possible a basic relay structure is chosen, 

 carefully planned for low maintenance effort, all of whose basic parts 

 can be made by mass production methods. On this basic framewoi'k one 



