A File Partitioning Model.
Abstract
The problem of reducing the average amount of data brought into core is considered for requests to a file system. A model is developed in which each request accesses one or more fields, each of whose length and probability of access is assumed to be known and independent of every other. The technique of partitioning the master file into subfiles by dividing the master record into subrecords is discussed and analyzed. An algorithm is presented to find an optimal partition for two special cases, and a heuristic for the general case is suggested. Finally, a formulation of the general problem as a zero-one integer programming problem is put forth. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 1972
- Accession Number
- AD0746494
Entities
People
- Stephen R. Kennedy
Organizations
- California Institute of Technology