TRUNK KEQriHI'^MI>:NTS IN ALTKHNATK KOUTINCi NKTWORKS 2!l7 



loads was arriveil at through an averaging? process. The ratio values are 

 <'oiiser\ative, i.e., there was some evi(leiic(> lliat in large toll offices the 

 higher rnlio values apply to larger loads than iiKlicated by the curve. 

 Il(»\ve\('i\ no attempt was made to develop separate sets of ratios for 

 ollices of different size since the use of the common set of values would 

 tend merely to increase slightly the num])er of HU trunks required in a 

 r(^lati\-ely few groups. The elimination of this small distortion did not 

 appear to warrant th(> effort recjuired to a('hie\-e it. The stud}' data also 

 showed that for a gi\en toll office the sum of the respective group busy- 

 hour loads for all groups was approximately 10 1 1 per cent greater than 

 the sum of their respective loads in the office bus\'-hour. 



The significance of this difference in the GBH and OBH aggregates is 

 at once apparent. The intertoU trvnik network designed on the alternate 

 routing principle is required to handle from 10-14 per cent less busy-hour 

 traffic than is required under the arrangement in which each toll center- 

 toll center load stands alone and must be trunked for its own busy-hour. 

 It is the pooling of trunk group capacities during the hour of maximum 

 traffic flow for a given office plus the increased efficiency due to larger 

 size of final groups that result in a retiuirement of fewer trunks in the 

 network as a whole than would be required with any non-alternate rout- 

 ing trunking system of comparable service characteristics. The alternate 

 routing system will enable the handling of normal busy-hour traffic on 

 \-irtuall3^ a no-delay basis in so far as trunk provision may be controlling. 

 The matter of speed of service potentialities and relative costs of alternate 

 routing ^'el■sus non-alternate routing systems will be treated later. 



SOME UNANSWERED QUESTIONS 



There are three ciuestions upon the answers to which depend ultimate 

 judgment of the adequacy and economy of a nationwide toll dialing net- 

 work constructed upon the principles and with the techniques already 

 described. These are: 



1. What will be the effect upon trunk requirements of a proper evalua- 

 tion of the non-random characteristic of lost calls, i.e., of the calls over- 

 flowed from high usage groups to other high usage or to final groups? 



2. What will be the effect upon trunk requirements of a proper evalua- 

 tion of the effect of non-coincidence of busy hoiu's and busy seasons 

 among the various toll centers?^ 



3. What are the relative net costs of a nationwide intertoll dialing 

 system engineered for multi-alternate routing and one designed without 

 engineered alternate routing? 



