OLAP (Online Analytical Processing) systems emerged in the early 1990s to address the analytical processing needs of relational databases, which were optimized for transactional workloads. OLAP servers introduced multidimensional data structures known as "cubes" that enabled businesses to perform complex queries swiftly and facilitated advanced data exploration techniques. However, OLAP had its challenges, including scalability issues due to its in-memory architecture, which made it hard and expensive to scale. As a result, alternative technologies like MPP databases and Hadoop emerged, leveraging distributed computing and flexible data processing models. The modern vision of a universal semantic layer aims to extract analytics modeling and aggregations from the BI layer and make them standalone, avoiding duplication across data and visualization tools in an organization. This architecture offers benefits, but also raises questions about alternative performance optimizations and communication protocols.