294 THE BELL SYSTEM TECHNICAL JOURNAL, MARCH, 1954 



to the TC2-SC group, then, would consist of the sum of the following 

 items: 



TC2-TC5, TC2-TC6, TC2-TC7, TC2-TC8, and TC2-SC 



Later examination of HU group possibilities with TC5 as the reference 

 TC would have to recognize that the TC2-TC5 item was already com- 

 mitted to a switched route via SC. Thus the TC5 summation would con- 

 sist of but four items (TC5-TC1, TC5-TC3, TC5-TC4, and TC5-P0) 

 instead of five, thereby reducing the possiblity of establishing a group 

 between TC5 and PO. Even the four-item summation might prove in a 

 group but if it be assumed further that a group proved in between TC3 

 and SC in addition to the TC2-SC group, the subsequent TC5 summation 

 to the PO area would be reduced to 3 items, namely, TC5-TC1, TC5-TC4 

 and TC5-P0. It is clear that by using the TC's homing on PO as refer- 

 ence points the tendency is to build up the HU groups to SC and dimin- 

 ish the possibility of establishing HU groups terminating at PO with the 

 result that SC handles more and PO less switched traffic. By thus select- 

 ing TC's homing on the lower class CSP, the PO, as the first set of refer- 

 ence points, an effect of "putting all eggs in one basket" is created since 

 SC is already a larger switching office by virtue of its classiffication in 

 the switching pattern. Therefore, it was deemed advisable to reverse the 

 point of view in order to distribute the switched load more equitabley 

 among the various CSP's. To accomplish this each CSP was given an 

 index number which reflected roughly both its size (total originating toll 

 traffic) and its importance (classification) in the switching pattern. The 

 most important CSP thus determined was assigned No. 1, the next most 

 important No. 2, and so on throughout the hst. Thus by starting with 

 the TC's homing on the CSP with Index 1 as reference points, then mov- 

 ing to the TC's on Index 2 and so on the desired distribution of switching 

 facilities was achieved. 



GROUP BUSY HOUR VERSUS OFFICE BUSY HOUR TRAFFIC 



Thus far we have considered the theory of alternate routing and the 

 more important techniques used in applying it to the design of intertoll 

 trunk networks. Another very important aspect of the whole problem 

 involves a difference in load levels between any two toll centers during 

 different hours of the day. 



It has been the aim in engineering intertoll trunk groups without alter- 

 nate routing to provide a sufficient number of trunks in each group to 

 handle its traflfic at a stated speed of service in the average busy hour of 

 the busy season of the particular group. All significant traffic data for 



