From Relational to Cloud: A Deep Dive into the Evolution and Market of Databases

1. Database Types and DBMS Fundamentals

Databases sit atop hardware and software layers, serving industries such as government, finance, energy, education, and transportation. They are broadly classified into relational databases (tables organized by rows and columns) and non‑relational (NoSQL) databases , which prioritize scalability and high read/write performance for massive, diverse data sets.

Relational Database Management Systems (RDBMS) store data in two‑dimensional tables, making querying intuitive via SQL. NoSQL databases emerged to address the limitations of traditional RDBMS in handling web‑scale, high‑concurrency workloads, offering easy horizontal expansion, simple data models, and superior performance on large data volumes.

A Database Management System (DBMS) is software that creates, uses, and maintains databases, providing functions such as data definition, manipulation, storage, maintenance, and multi‑user access. Its workflow includes:

Accepting data requests from applications;

Translating high‑level requests into low‑level machine instructions;

Executing database operations;

Retrieving query results;

Formatting results;

Returning the final output to the user.

DBMS implementations vary by data model—hierarchical, network, relational, object‑oriented—leading to differences in interfaces and capabilities across systems.

2. Historical Evolution of Databases

The development of databases can be divided into three major stages plus a recent dynamic:

Navigational databases (1960s) : Early systems like IDS and IBM’s IMS provided pointer‑based access paths, limited by hardware constraints.

Relational databases (1970s‑present) : Edgar Codd introduced the relational model; IBM’s System R and DB2, Oracle, Ingres, PostgreSQL, MySQL, and Microsoft Access followed, establishing SQL as the standard query language.

Non‑relational (NoSQL) and NewSQL (2000s‑present) : The term “NoSQL” was coined in 1998 and revived in 2009 to describe distributed, schema‑flexible stores such as DynamoDB, MongoDB, Cassandra, and Redis. NewSQL combines the scalability of NoSQL with the ACID guarantees of relational systems.

Since 2017, the concept of a self‑driving (autonomous) DBMS has emerged, leveraging AI and machine learning to optimize workloads, predict resource needs, and automate administration, though the technology remains in early stages.

3. Market Size and Growth Potential

According to Gartner, the global database software market was $46.1 billion in 2018 and is projected to reach $54.9 billion by 2021, growing at a 9.1% CAGR. In China, the market expanded from ¥35.03 billion in 2009 to ¥149.91 billion in 2018, with an estimated ¥200 billion by 2020 and a 17.86% annual growth rate.

Foreign vendors still dominate, holding roughly 60% of the global market (Oracle 40.1%, IBM 11%, Microsoft 8.3%). Domestic Chinese vendors collectively hold about 14% of the Chinese market, with leaders such as Nanda General and Wuhan DM (18% and 13% respectively) and smaller players each below 5%.

4. Leading Vendors and Industry Rankings

DB‑Engines ranks database popularity monthly; as of March 2020, Oracle leads globally, followed by MySQL and Microsoft SQL Server. In China, IBM DB2 ranks sixth, while SAP HANA and Adaptive Server appear in the top twenty.

Gartner’s Magic Quadrant (2018) lists Amazon, Microsoft, Oracle, SAP, and IBM as leaders. Alibaba Cloud entered the Visionaries quadrant in 2019, becoming the first Chinese vendor to appear, but domestic vendors still lag behind global leaders.

5. The Cloud Migration Trend

Gartner predicts cloud databases will capture 50% of the market by 2021 and 75% by 2023. Cloud databases deliver high scalability, availability, multi‑tenant resource distribution, and simplify hardware/software provisioning.

From a data‑model perspective, cloud databases are not a new technology; they simply offer database services (relational or NoSQL) as a managed cloud offering, combining the elasticity of cloud computing with the familiarity of traditional database interfaces.