Research and Comparative Analysis of Approaches to Data Storage Architecture, Hashing, Indexing, and data Compression in Oracle and PostgreSQL

The paper presents a comparative analysis of the Oracle and PostgreSQL database management systems based on key architectural criteria: data storage, hashing mechanisms, indexing, and data compression. The aim of the research is to identify architectural differences and assess their impact on performance and scalability. It is shown that Oracle implements a complex multi-level storage model (tablespace–segment–extent–block), providing efficient space management and high data access speed under various workloads. PostgreSQL, in contrast, employs a simpler page-oriented structure (8 KB per page), which simplifies administration but reduces efficiency in analytical data processing. When comparing hashing mechanisms, it is established that Oracle, through the Hash Cluster feature, accesses the required data block directly (on average 0.5 ms per query), whereas PostgreSQL performs an additional table lookup after indexing, which approximately doubles access time. Both DBMSs utilize B-tree indexing with logarithmic search complexity; however, Oracle additionally employs specialized index types (Bitmap, IOT), while PostgreSQL focuses on extensibility and support for user-defined structures (GiST, GIN, BRIN). In terms of data compression, Oracle achieves compression ratios of up to 8–10× (using HCC and In-Memory Column Store) without performance loss, whereas PostgreSQL provides 3–5× compression through TOAST and external columnar extensions. The comparative analysis demonstrates that Oracle is oriented toward performance and scalability, while PostgreSQL emphasizes flexibility and architectural extensibility.