Performance evaluation using equivalent batch-call traffic models
Zenji Hayashi, Satoshi Nojo · Electronics and Communications in Japan (Part I Communications) · 1997
Traffic evaluation for non-random calls becomes more important as networks move toward integrated services digital network status. However, the conventional evaluation methods have many problems: the evaluations are limited to Poisson call distributions because of difficulty in analysis; the evaluation methods are very complex; and they take a lot of time to compute as the number of systems increases. Conventionally, non-random call traffic is evaluated by developing a method for each system. The generally applicable evaluation method approximates arrival calls in terms of batch random calls. This method is effective for evaluating overflow calls, but it is not always effective for general non-random calls. In this situation, this paper proposes an improvement over the conventional method. The improvement is that the parameters of the approximate model are evaluated in terms of batch calls at the arrival time of the traffic under evaluation. The proposed method gives good precision which will be shown by an example of a fully available trunk-group. It is simple, allows rapid calculation, and is effective for evaluating many systems. © 1997 Scripta Technica, Inc. Electron Comm Jpn Pt 1, 80(12), 90–99, 1997