298 THE BELL SYSTEM TECHNICAL JOURNAL, MARCH 1953 



presented here only to illustrate the application of throwdown tech- 

 niques. However, as problems become more complex it becomes in- 

 creasingly difficult to apply analytical methods. Even in relatively simple 

 systems the interplay of variables may be so involved that existing theo- 

 retical methods are entirely inadequate. Various simplifying assump- 

 tions must be made and there is often the doubt that some important 

 factor has not been overlooked in formulating the mathematical theory. 



The most fruitful use of throwdown methods has been to check results 

 obtained by theory and to obtain data upon which mathematical theories 

 can be based. Throwdown techniques, of course, can be used to obtain 

 direct results, but the functioning of a system will be better understood 

 if there is at least some attempt to develop a theory which explains how 

 various forces act together to produce observed results. Such theories 

 may suggest modifications of the system which will improve its per- 

 formance. 



It can be seen that the planning of a throwdown study requires a 

 thorough knowledge of both the functioning of the system being simu- 

 lated and the characteristics of the input data being processed by this 

 system. The validity of results will depend upon the faithfulness with 

 which the artificially prepared input data represent real data and the 

 accuracy with which the throwdown routines represent the real system 

 performance. The following two sections of this paper will describe the 

 No. 5 crossbar system and the characteristics of the subscribers using 

 this system. Later sections describe the methods used in the machine 

 for simulating the dynamic performance of an operating system with 

 its subscribers. 



THE NUMBER 5 CROSSBAR SWITCHING SYSTEM* 



No. 5 crossbar is a marker-controlled system designed primarily for 

 local central office application in the residential sections of large cities 

 and the fringe areas surrounding these cities. 



In regions of this type a relatively large proportion of all calls are 

 completed to subscribers within the same office. Since the surrounding 

 offices to which connection must be made are likely to be of widely 

 diversified types, the system is designed to interconnect with any existing 

 type of central office. No. 5 is also capable of serving isolated centers 

 from about 3,000 lines up, and multioffice areas including the largest 



*F. A. Korn and James S.Ferguson, Trans. A.LE.E., 69, Part 1, pp. 244-254, 

 1960. 



