September 2018 Summaries

Multi-tiered database architectures, which incorporate various database types like NoSQL, SQL, and Object stores, have gained popularity due to their ability to handle scenarios such as data streaming, where large volumes of unstructured data require persistent storage. Despite their advantages, these architectures pose significant challenges including data integrity issues, complexity, and synchronization difficulties. Faults in synchronization processes can lead to databases being out of sync, undermining trust in data accuracy, especially in real-time reporting contexts, and are further complicated by the need for mono-directional data flow to avoid corruption. The complexity of these systems often necessitates extensive knowledge that can be problematic if key personnel are unavailable, as illustrated by the author's personal experiences. To address these issues, newer databases adopting a Hybrid Transactional/Analytical Processing (HTAP) model are suggested, allowing for the management of operational, transactional, and analytical workloads within a single system, thus eliminating the need for multiple databases and mitigating the associated risks.