Packet Switch Provisioning: Accounting for Uncertainty in the Forecasting of Customer Characteristics

Frequently emerging switch services are not sufficiently well understood to be accurately forecast. Since there is a great deal of uncertainty in the forecast, accurate engineering of the switch is extremely difficult. Distributed networks of packet switches are particularly affected by these uncertainties, because the resources available to individual users are shared and limited. In this paper we demonstrate a methodology for accounting for the effects of uncertainty in the forecasting of customer characteristics in packet switch provisioning. We show that depending on the degree of uncertainty in estimating parameters for the arrival process, the application of the methodology can have a substantial impact on the provisioning of the switch.