How OceanBase Overcame the Odds to Become China’s Homegrown Distributed Database

OceanBase is a fully self‑developed distributed relational database by Alibaba and Ant Financial, the world’s first such system used in core financial services. Development began in June 2010, and from the start the project faced severe difficulties, even nearing team dissolution when no business would adopt it.

Eventually OceanBase crossed the “valley of death”, fully replacing Oracle within Ant Financial and handling the massive load of Double 11 for five years, achieving a peak of 256,000 payments per second and 42 million requests per second. Since 2017 it has been commercialized, being deployed in dozens of commercial banks such as Nanjing Bank, Zheshang Bank, Suzhou Bank and PICC Health Insurance, where it increased loan‑transaction capacity tenfold and cut per‑account costs from 30‑50 CNY to 4 CNY, processing millions of loans per day with average response time under one second.

According to senior researcher and OceanBase founder Yang Zhenkun, the success was inevitable given industry and era conditions.

Opportunity

In 2009 a wave of NoSQL databases emerged, sparking a revolution that questioned the relevance of relational databases. However, the rise of cloud computing created new demands for large‑scale databases, exposing the limitations of traditional relational systems—poor scalability, limited capacity, high cost—making many believe relational databases were on the brink of extinction.

Yang saw a “once‑in‑a‑lifetime” chance: cloud computing required massive concurrency and data volume, and Alibaba’s own services (Taobao, Alipay) relied heavily on relational databases that could not meet these new requirements.

Choice

“Most people are smart, but few can turn that intelligence into results. Success comes from focusing wisdom on what must be done, not on what is merely clever.”

Yang, a mathematics graduate of Peking University who later studied under computer‑science pioneers, chose to build a distributed relational database from scratch rather than follow existing solutions. He left a comfortable technical management path at Taobao after only two weeks to assemble a new team and create OceanBase.

Building a distributed relational system meant tackling fundamental challenges: ensuring high availability and data consistency together, achieving CAP requirements, and replacing costly hardware‑based reliability with software‑driven solutions.

Initially, the project received little support; many doubted its feasibility. Yet the lack of attention gave the team years to mature the technology in relative obscurity.

Setback and Revival

By 2012 the team was transferred to Alipay, where the financial domain presented even greater database challenges. In the summer of 2013 Alipay launched its “remove IOE” initiative, seeking to replace Oracle. Yang proposed OceanBase’s solution, which introduced a “one‑master‑multiple‑backup” model that could tolerate failures while maintaining consistency.

During the 2014 Double 11 shopping festival, OceanBase handled 10 % of transaction traffic, later increasing to 10 % of total traffic and earning the team the 2015 Ant Financial CEO award.

From 2015 onward, OceanBase replaced Oracle in Alipay’s transaction, payment, and accounting systems, and by 2017 it began commercial deployments in commercial banks.

Changing Landscape

While NoSQL gained prominence a decade ago, recent years have seen a resurgence of SQL‑based services from cloud providers (Amazon Aurora, Google Cloud SQL, Azure PostgreSQL). Gartner predicts relational databases will power at least 70 % of new applications by 2020.

Today, OceanBase aims to achieve seamless Oracle compatibility, allowing migrations without code changes. Yang believes the future of data warehousing will converge on distributed relational systems like Google Spanner, and OceanBase will continue to evolve faster and farther.