Enhancing the B+-tree by dynamic node popularity caching

  • Authors:
  • Cui Yu;James Bailey;Julian Montefusco;Rui Zhang;Jiling Zhong

  • Affiliations:
  • Monmouth University, USA;University of Melbourne, Australia;Monmouth University, USA;University of Melbourne, Australia;Troy University, USA

  • Venue:
  • Information Processing Letters
  • Year:
  • 2010

Quantified Score

Hi-index 0.89

Visualization

Abstract

We propose the Cache Coherent B^+-tree (CCB^+-tree), an indexing structure that can improve search performance compared to the traditional B^+-tree. The CCB^+-tree makes use of the unused space in the internal nodes of a B^+-tree to cache frequently queried leaf node addresses, thus saving node accesses. An experimental study shows that the CCB^+-tree can outperform the traditional B^+-tree on workloads where certain queries are much more popular than the others.