280 THE BELL .SYSTEM TECHNICAL JOURNAL, MAKCH, 1954 



groups carrying traffic in only one direction. The dashed lines represent 

 the direct triuik groups between central offices and will be referred to as 

 "high usage" groups. The solid lines represent trunk groups in the alter- 

 nate route and will be referred to as "final" groups since they constitute 

 the last choice path for traffic to N. It will be seen that the groups AN, 

 BN, and CN correspond roughly to the groups of individual trunks 

 serving a, b, and c in Fig. 1 and that group TN corresponds to the com- 

 mon trimks. But here the analogy breaks down, for the traffic from A, 

 B, and C, respectively, to X will rarely be equal and there is no counter- 

 part in Fig. 1 to the groups from A, B, and C to T. At this point it 

 becomes clear that the technicjues for determining the capacities of 

 simple graded multiples are of no avail in dertemining the proper number 

 of trunks in each group of such a system. The problem is further com- 

 pficated by the fact that offices A, B, and C will offer to their respective 

 tandem groups traffic for which such groups are the only route. Such 

 traffic would consist of a number of relatively small items destined for 

 other offices, not indicated in Fig. 2, and for which direct routing would 

 be uneconomical. 



To make such a system work there had to be available at the originat- 

 ing office a switching mechanism capable of testing two or more triuik 

 groups in succession for a single call. The Xo. 1 local crossbar office 

 developed in the 1930's pro\ided such facilities and a trial of the alter- 

 nate routing scheme for trunking between local offices was made at two 

 New York City offices in 1941. 



The basic aim was to so arrange the trunking layout at each office 

 that the traffic to every other office in the city would be carried as eco- 

 nomically as possible over some combination of a\'ailable routes. This 

 objective reciuired determination of the cost of a direct route from the 

 trial office to every other office and also the cost of any potential alter- 

 nate route. The relationship between the costs of the several routes 

 available to a giA-en item, not the absolute values of the facilities was 

 the important factor. It is not the intent here to discuss the derivation 

 of such factors but rather to describe how they were used in determining 

 the interoffice trunk layout. 



The end result of the examination of line and switching costs for the 

 several routes was a ratio of the cost of each potential alternate route 

 to the cost of the direct route to each distant office. The smallest of such 

 ratios would determine the alternate route to be used unless the ratio 

 were unity or less, a rare circumstance which would indicate that no 

 direct trunks should be pro\-ided. The next step in the problem was to 

 provide trunks in such numbers and arrangements that the cost of 



