Concurrent search structure algorithms

  • Authors:
  • Dennis Shasha;Nathan Goodman

  • Affiliations:
  • New York Univ., New York, NY;Kendall Square Research Inc., Cambridge, MA

  • Venue:
  • ACM Transactions on Database Systems (TODS)
  • Year:
  • 1988

Quantified Score

Hi-index 0.01

Visualization

Abstract

A dictionary is an abstract data type supporting the actions member, insert, and delete. A search structure is a data structure used to implement a dictionary. Examples include B trees, hash structures, and unordered lists. Concurrent algorithms on search structures can achieve more parallelism than standard concurrency control methods would suggest, by exploiting the fact that many different search structure states represent one dictionary state. We present a framework for verifying such algorithms and for inventing new ones. We give several examples, one of which exploits the structure of Banyan family interconnection networks. We also discuss the interaction between concurrency control and recovery as applied to search structures.