The Cloud Era of Databases: Insights from OceanBase Chief Scientist Yang Zhenkun

At the OceanBase 2nd Developer Conference, chief scientist Yang Zhenkun delivered a keynote titled “Databases in the Cloud Era,” emphasizing that the core characteristic of the cloud is resource pooling, sharing, and reuse, which will soon allow distributed databases to fully leverage pooled compute and storage resources.

He reviewed the historical development of databases: commercial databases in the 1980s lacked robust transaction support; the 1990s saw the rise of open‑source databases that matured over a decade; the internet boom of the 1990s introduced new demands for massive concurrency, scale, and cost‑effectiveness.

Despite this evolution, many newer systems still lack strong transaction processing capabilities, leaving relational databases as the only proven production‑grade solution for ACID transactions.

He used the analogy of a large aircraft versus many small aircraft to illustrate why native distributed databases, rather than merely scaling single‑node machines, are essential for truly elastic database capacity in the cloud.

The cloud’s fundamental traits—resource pooling, sharing, and reuse—dramatically improve CPU utilization (e.g., raising average usage from 3‑5% to 30%), delivering six‑fold cost efficiency and enabling shared DBA expertise across many enterprises.

Cloud databases must also support diverse data modalities—semi‑structured and unstructured data—while providing on‑demand, pay‑as‑you‑go compute and storage, exemplified by scenarios where a user may need petabytes of storage with a single CPU core or millions of CPU cores for modest data volumes.

These examples illustrate how both compute and storage capacities can be pooled and shared in the cloud, just like any other cloud resource.

He concluded by reaffirming that cloud‑native databases will become an integral part of the cloud ecosystem, offering fully shared, elastic, and cost‑effective services.