Analysis and comparison of data structures for database indexing

Abstract

The consecutive evolution of a digital communication, remote data and services provision and interconnection and cloud technologies environment requires the development and construction of efficient database management systems (DBMSs) that implement storage, distribution and management of large datasets. These DBMSs are composed of multiple file sub-systems and system structural components layers that implement the efficient and fast management of the stored datasets through transnational operations sets. Indexing systems have been developed in the context of DBMSs in order to implement this functional efficiency, operating as file sub-systems that are structurally and functionally connected to a large and complex set of other database sub-systems. These systems are also composed of multiple mechanisms comprising indexing data structures sets and algorithms. This thesis composes an implementation of the B-tree, B+-tree, B-Hash Map and B+-Hash Map index structures and a complete theoretical and computational analysis of those structures functional and structural levels. The computational analysis, evaluation and comparison of the developed B-tree, B+-tree, B-Hash Map and B+-Hash Map index structures functional efficiency and time performance was carried out through a set of computational processes on real and synthetic datasets. Consequently, this study provides a quite complete, comprehensive and detailed theoretical and computational analysis of these implemented - developed indexes structure and functions performance in terms of time and memory resources management efficiency and time performance.