• Enterprise analytics centered on data warehouses emphasized precomputation and schema design: conceptual multidimensional modeling and view selection to materialize aggregates, enabling predictable query performance over large integrated stores [1], [8], [13], [14].

• Sublinear-space approximation became a core analytic strategy for massive datasets and streams, using synopsis data structures and randomized sketches to estimate norms and query answers in one pass with controllable error [5], [6], [15], [17].

• Big Data shifted from batch mining to continuous monitoring, framing streams as first-class managed data with standing queries, online analytics, and windowed operators, supported by streaming algorithms for low-memory updates [3], [5], [17].

• Scalable knowledge discovery prioritized algorithmic redesign—parallelization, data partitioning, and index-aware computation—to mine large datasets: parallel decision-tree classification, grid-based clustering, fast principal component analysis, and distance-based outlier detection [4], [7], [9], [10], [11], [19].

• Systems work pushed computation toward the data via specialized middleware, archival storage interfaces, and access-control and privacy-aware data governance, enabling in-situ filtering of scientific archives and controlled access to large indexes and linked datasets [2], [16], [20].