The interval B-tree

Abstract

The time index [4] is an index structure built to speed up the access of data with valid times satisfying some given conditions. Unfortunately its space complexity is O(n2) where n is the number of time intervals being stored. As a result, high disk I/O overhead is experienced in various operations such as the insertion and deletion of a time interval (O(n)) and the interval intersection operation (O(n2)). A new index structure is proposed using a B+-tree as the underlying structure. Its space complexity is reduced from O(n2) to O(n). The time complexity of insertion and deletion of an interval is reduced from O(n) to O(log n), and that of interval intersection is reduced from O(n2) to O(log n + F) where F is the time used to report intersections. © 1995 Elsevier Science B.V. All rights reserved.