Author(s): Coffman EG, Randell B
Abstract: This paper concerns the problem of obtaining predictions of the extend to which additional core storage would improve the performance of a given paging system based on information that could be obtained from monitoring the system whilst running its normal workload. It is shown that for a large class of replacement algorithms there are efficient techniques for producing exact predictions of the performance improvement, and that for a further class of algorithms statistical predictions can be provided.
|
Professor Brian Randell
|
|