An optimal external selection algorithm and its application in the Internet

This paper presents an optimal sampling external selection (SES) algorithm to select k-th smallest item in large data sets for the two-level memory model. Based on the SES algorithm, two algorithms SES/DASK (dynamic assigned sorting key) and SES/FASK (fixed assigned sorting key) are also proposed which are applied to the worldwide selection problem in the Internet environment. We use the sampling information scheme to form an elegant and simple algorithm to reduce the number of disk I/Os. Especially, our algorithm is more efficient for the multiple selections.