300 THE BELL SYSTEM TECHNICAL JOURNAL, MARCH, iDol 



soundness of the whole project. In the early stages of planning the l)ases 

 for making such a judgment were rather meager since the only actual 

 experience with alternate routing trunking had been in local systems, 

 notably in New York City. Furthermore, the cost of the complicated 

 switching mechanism, known to be essential to the plan but not yet 

 developed, was an estimate of uncertain reliability. Added to these were 

 the ever-present hazards of estimating the volume and distril)ution of 

 toll traffic some 10 to 15 years in the future. However, in the light of the 

 information then available it appeared that the overall cost of the proposed 

 arrangments would be approximately equivalent to the cost of continu- 

 ing present methods of engineering geared to a satisfactory trunk speed 

 of service in the busy hour. Busy -hour trunk speed is defined as the aver- 

 age interval of delay* experienced in the busy hour by an attempt at 

 trunk seizure. This interval excludes any effect of operating procedure 

 as such and hence reflects only the adequacy of the trunk plant to handle 

 the offered load. It should be noted parenthetically that the use of busy 

 hour trunk speed as a criterion of adequate intertoU trunk provision is 

 different from the criterion used for interlocal trunk provision. The 

 former refers to the duration of delay encountered by the average at- 

 tempt at trunk seizure whereas the latter concerns the percentage of 

 attempts that will encounter delay of any duration. These different 

 service criteria arose from the traditional need for high speed handling 

 of large volumes of local traffic and the acceptability of a much slower 

 speed for toll traffic. 



The reference above to a "satisfactory" trunk speed of service pre- 

 sumes a stated objective to be achieved under the current method of 

 engineering. In 1945 the service objective was expressed in different terms 

 which can be interpreted as being roughly equivalent to a busy-hour 

 trunk speed of 30 seconds. It is obvious that a nationwide network de- 

 signed to produce an average busy-hour trunk speed, of, say 20 seconds 

 would require more trunks than one designed to produce an average 

 delay of 30 seconds. With an alternate routing system, however, there 

 is only one speed of service under normal load conditions, namely, one 

 in which the average delay is so small as to be incapable of meaningful 

 expression in terms of seconds per average attempt. It follows then that 

 in any attempt to assess the comparative costs of outside plant and 

 equipment luider the ciu'rent method with those under the alternate 

 routing method of engineering trunks it becomes necessary to first define 

 the service level which the former is required to satisfy. Objective service 



* S. C. Rappleye, A Studv of the Delays Encountered by Toll Operators in 

 Obtaining an Idle Trunk, B.S. T. J., 25, Oct., 1946. 



