Cost Model Assumptions

Contents

Cost Model Assumptions#

  • B: Number of data blocks in the file

  • R: Number of records per block

  • D: Average time to read/write disk to block

    • Ignore difference between sequential vs. Random I/O

Assumptions#

  • Average case for uniform random workloads

  • Ignore:

    • Sequential vs. Random I/O

    • Pre-fetching

    • Any in-memory cost

Further strict assumptions:

  • Single record insert and delete

  • Exactly one match for equality

  • Heap files always append at end of file

  • Sorted files are packed and sorted according to search key

Contents