Parallel Select

Abstract

Chapter 18 introduced the concept of an inverted index to prevent the database system from performing a full column scan every time a query searches for a predicate in the column. However, maintaining an index is expensive and consumes additional memory. So the decision to use an index should be made carefully, balancing all the pros and cons an index would bring in the particular situation. This chapter discusses how to speed up a full column scan despite adding an index to it. Chapter 17 introduced parallelism as a means to parallelize the execution of database operations. In this chapter, we present a detailed description of how parallelism can be used to speed up the execution of a SELECT operation.

Self Test Questions

 

• 1. Amdahl’s Law

  Amdahl’s Law states that...

  • (a) the number of CPUs doubles every year

  • (b) the level of parallelization can be no higher than the number of available CPUs

  • (c) the speedup of parallelization is limited by the time needed for the sequential fractions of the program

  • (d) the amount of available memory doubles every year.

• 2. Query Execution Plans in Parallelizing SELECTS

  When a SELECT statement is executed in parallel...

  • (a) all other SELECT statements are paused

  • (b) its query execution plan becomes much simpler compared to sequential execution

  • (c) its query execution plan is adapted accordingly

  • (d) its query execution plan is not changed at all.