Meet the Database Built for Modern-Day Pressures (And Your AI Dreams Too)

Ever wondered how China's biggest online shopping event, Double 11 (Singles' Day), doesn't melt down under the strain of millions of simultaneous transactions? The answer lies in a database called OceanBase.

Now, the namesake company is thinking big. It sees the advantages of bringing its database's ability to handle Alipay's extreme demands to the global financial sector, drowning high data loads, volumes and real-time transactions. And it is correct.

The origin story: Born from a scaling crisis

To understand OceanBase, we must first dial back to 11 years ago. Alipay, already a major third-party payment platform, was creaking at its seams. The problem: it's a monolithic Oracle database.

Like Oracle's monolithic architecture, legacy systems couldn't cope with the data volume and request load. They simply weren't built to scale or gracefully handle multi-region disaster recovery.

But it was not the only concern. "The technical challenges weren't just about the database," explains Ted Bai, head of solution architecture for the international business at OceanBase. "The entire infrastructure was a massive fabric under extreme stress."

For Alipay to dream big, it had to invest more money in hardware and database upgrades. It knew it had to go back to the drawing board. The solution that emerged was OceanBase.

Built from the ground up as a fully distributed database, it shattered performance bottlenecks and enabled seamless scaling.

This wasn't just theory, either. In 2019, OceanBase made history as the first distributed database to ace the TPC-C benchmark with its V2.2 entry, beating out Oracle using standard x86 servers instead of high-end hardware with a score of 60 million tpmC. In 2020, it set a new record with 700 million tpmC.

OceanBase was the first distributed database to enter the coveted TPC-C rankings, often seen as a Formula 1 race for databases, and the first Chinese database to do so. The leap in performance also highlighted the power of distributed architecture in handling high-volume financial transactions.

The stranglehold of the monolithic databases was now splayed open.

Open source, but not quite forked

Since becoming Alipay's database of choice, OceanBase has grown into a commercial product. It now serves various industries, from banks and insurers to telecom giants and even the Chinese government.

The architecture has matured to meet the unique needs of each sector, including cloud-based deployments on platforms like AWS and GCP.

OceanBase is open source, a significant difference from the rest of the TPC-C competitors. This decision has been crucial for broader adoption.

"To make it a truly global product, it had to be open source," Bai says. This fosters community engagement and accelerates innovation, while the enterprise version focuses on Oracle compatibility and advanced security features, catering to the rigorous demands of financial institutions.

Where OceanBase excels

Ask any database engineer, and they will tell you that performance alone is insufficient. It also had to be secure, cost-efficient and flexible.

For financial institutions, security isn't a feature; it's table stakes. OceanBase offers transparent data encryption (TDE SSL), granular access controls, and comprehensive audit trails. Its SOC2 and PCI DSS certifications prove it means business, adds Bai.

The bottom line also matters. OceanBase’s multi-tenancy design lets you consolidate multiple databases into a single cluster, maximizing resource utilization, explains Bai. Add to that an innovative storage engine with impressive compression rates, and you've got a recipe for significant cost savings.

Lastly, OceanBase caters to a wide range of deployment scenarios, whether you're fully on-premises, cloud-native, or somewhere in between. This multi-infrastructure approach means you're not locked into a single vendor, allowing you to choose what works best for your organization.

These features are helping OceanBase make significant inroads across the financial services sector, from established insurers to leading-edge digital banks and Web3 companies.

Why customers are switching

Just ask China Life Insurance, one of the biggest insurers in China.

They migrated hundreds of Oracle database clusters to OceanBase, driven by performance bottlenecks and ballooning costs.

Another primary reason was that they were refactoring their next-generation applications to follow a distributed and microservice design. OceanBase's native distributed architecture made it the ideal choice.

OceanBase's seamless Oracle compatibility through features commonly used in the latter made this transition remarkably smooth.

Bai also noted that digital banks are beginning to deploy OceanBase as they shift toward a distributed architecture. The database's ability to perform rolling upgrades was a major plus compared to traditional databases, which need maintenance shutdowns.

Meanwhile, Bai shared that the company works closely with a Web3 company because of scalability and reliability. In particular, the optimized use of Multi-Paxos allows it to implement multi-replica data synchronization and cluster high availability.

Essentially, Multi-Paxos helps the company avoid split brains: when a network failure at the primary node causes a company to make a disaster recovery decision to spin up the standby node, leading to multiple writes and reduced data correctness.

The road ahead

OceanBase isn't done with its journey. According to Bai, it’s just starting.

Their roadmap includes developments like full-fledged Hybrid Transactional/Analytical Processing (HTAP) capabilities, meaning you can run transactional workloads and complex analytics within the same database. Bai also noted that the company is venturing into the realm of NoSQL with their KV store, providing a high-performance alternative to systems like HBase.

However, the truly groundbreaking development is integrating vector database plugins for generative AI applications, positioning OceanBase not just as a data store but as a powerful engine for FSI's AI-driven future.

Let Ted Bai and leading FSI companies explain how OceanBase is reshaping the future of financial data management one dataset at a time. Join our upcoming luncheon "Decoding the Data Advantage: How Modern Databases Drive Financial Innovation" on July 10, 2024. To register, email [email protected].

Image credit: iStockphoto/Shutthiphong Chandaeng