Optimal Dynamic Range Searching in Non-replicating Index Structures

  • Authors:
  • Kothuri Venkata Ravi Kanth;Ambuj K. Singh

  • Affiliations:
  • -;-

  • Venue:
  • ICDT '99 Proceedings of the 7th International Conference on Database Theory
  • Year:
  • 1999

Quantified Score

Hi-index 0.01

Visualization

Abstract

In this paper, we examine the complexity of multi-dimensional range searching in non-replicating index structures. Such nonreplicating structures achieve low storage costs and fast update times due to lack of multiple copies. We first obtain a lower bound for range searching in non-replicating structures. Assuming a simple tree structure model of an index, we prove that the worst-case time for a query retrieving t out of n data items is Ω((n/b)(d-1)/d + t/b), where d is the data dimensionality and b is the capacity of index nodes. We then propose a new index structure, called the O-tree, that achieves this query time in dynamic environments. Updates are supported in O(logb n) amortized time and exact match queries in O(logb n) worst-case time. This structure improves the query time of the best known non-replicating structure, the divided k-d tree, and is optimal for both queries and updates in non-replicating tree structures.