Efficient Storage for Analytics: Columnar Storage & Compression Techniques

This lecture covers storage technologies for analytics, focusing on columnar storage and compression techniques. It explains the benefits of Decomposition Storage Model (DSM), including DSM properties, page format, and write operations. The lecture also discusses compression methods such as lossless compression, run-length encoding, bit-vector encoding, and dictionary encoding. It explores how operators work over compressed data and introduces the concept of Write-Optimized Storage (WOS). Additionally, it delves into Partition Attributes Across (PAX) and its advantages, as well as the transition from DSM to PAX. The lecture concludes by emphasizing the importance of choosing the right storage layout based on the workload requirements.