Leveraging Networking Techniques to Improve Data Integration for Business Intelligence Systems

By Malvina NIKLEKAJ, Jora BANDA

Abstract
Integrating heterogeneous data sources for Business Intelligence (BI) systems remains a critical challenge in the digital transformation of organizations. Most of the literature focuses on traditional integration techniques such as ETL/ELT or on the use of advanced architecture such as data lakes and data warehouses, often neglecting the impact that network protocols and transmission policies have on the process. This study proposes a new “networking-aware” approach to data integration in BI, analyzing the role that protocols such as HTTP/2, HTTP/3/QUIC, gRPC, MQTT or WebSockets play in ensuring data freshness, reducing latency and managing network costs. The methodology followed includes analytical performance modeling through queuing theory (queuing models), formalization of metrics for measuring latency and data staleness, and analysis of typical integration scenarios (CRM–ERP, IoT streaming, and hybrid integration). The predicted results highlight the advantages and limitations of different protocols, suggesting optimal combinations for different use cases. The main contribution of this paper is the creation of a conceptual framework that links network metrics to BI integration, providing practical recommendations for architects and researchers. The conclusions show that choices at the network layer are not only technical issues of data transmission, but essential elements that directly affect the quality, consistency, and cost of business analysis.

Key words: Data Integration, BI, Network Protocols, CDC, ETL Process.

https://doi.org/10.58944/vigs8730