Aggregated Data in Tree-Based Index Structures

Abstract

This chapter describes an approach where aggregated data is materialized in the inner nodes of an index structure. These data could also be calculated from data stored in its successor nodes, but the access to aggregated data in the inner nodes is faster than accesses to successor nodes. This concept of aggregated data in the inner nodes is generic and is applicable to most treebased index structures. This chapter describes this concept of aggregated data in the inner nodes in detail, and it shows how the algorithms are modified to maintain and use the aggregated data. At the end of this chapter we present results of experiments where the extended structure with aggregated data in the inner nodes is compared with the standard structure without materialized aggregated data.