HORIZONTAL FRAGMENTATION AND SELECTIVE REPLICATION IN DISTRIBUTED DATABASES: A PROPOSED ARCHITECTURAL MODEL FOR OPTIMIZING THE E-SISTAFE SYSTEM IN MOZAMBIQUE

Abstract

In an increasingly digitalized world, the operational efficiency and responsiveness of information systems are crucial for decision-making in organizations, including governmental ones. The choice of database architecture, between centralized (client-server) and distributed, is a determining factor for performance and scalability. This article proposes an analysis of the migration of Mozambique's public financial management system, e-SISTAFE, from a client-server architecture to a distributed database architecture with a focus on horizontal fragmentation. The objective is to demonstrate, through an in-depth literature review and the application of a Design Science Research (DSR) model for artifact proposition, that fragmentation can mitigate performance bottlenecks, network overload, and latency in geographically dispersed systems. The literature review focuses on the challenges of consistency, transparency, and query optimization inherent to fragmentation, and integrates recent literature on Distributed Databases (DDB) and governmental system migration. It is concluded that, although fragmentation requires a substantial investment in infrastructure and qualified human resources, its adoption in e-SISTAFE has the potential to significantly improve the efficiency of expenditure execution and revenue management at the national level.