OceanBase Database is a native, enterprise-class distributed database developed independently by the OceanBase team. It provides financial-grade high availability on commodity hardware and introduces a new city-level disaster recovery standard known as "Five IDCs across Three Regions". As the world's first distributed database to pass the TPC-C benchmark test, it can support over 1,500 nodes in a single cluster. Additionally, it features cloud-native architecture, strong consistency, and high compatibility with other popular databases such as Oracle and MySQL.
Core features
Distributed transaction engine
The distributed transaction engine of OceanBase Database supports the atomicity, consistency, isolation, and durability (ACID) properties of transactions and provides strong data consistency in the entire cluster.
OceanBase Database creates multiple replicas for transaction logs based on Paxos. This ensures the availability and persistence of transactions.
Transparency and scalability
OceanBase Database provides innovative RootService and partition-level load balancing to ensure high scalability. OceanBase Database supports smooth online scaling. After scale-out, OceanBase Database automatically performs load balancing, which is transparent to applications and does not affect system continuity.
In addition, OceanBase Database can dynamically scale out an ultra-large-scale cluster that contains more than 1,500 nodes, more than 3 petabytes of data, and more than a trillion rows in one table. In the TPC-C benchmark test, OceanBase Database can achieve a scaling ratio of 1:0.9. This helps you maximize the return on your hardware investment.
High availability
OceanBase Database uses a shared-nothing architecture with multiple replicas to ensure zero single point of failure (SPOF) and system continuity. OceanBase Database supports high availability and disaster recovery at the node, IDC, and region levels. OceanBase Database can be deployed in a single IDC, two IDCs, three IDCs across two regions, and five IDCs across three regions. Test results show that OceanBase Database can reach a recovery point objective (RPO) of 0 seconds and a recovery time objective (RTO) of less than 30 seconds. This means that OceanBase Database has the level-6 disaster recovery capability, which is the highest level in the international standard for disaster recovery capabilities.
OceanBase Database supports primary and standby databases based on log replication to provide you with more flexible high-availability and disaster recovery capabilities. The primary cluster sends transaction logs to the standby cluster to synchronize data. This ensures that the application system can rapidly recover if the primary cluster encounters data damage or a disaster. When the primary cluster of an OceanBase Database becomes unavailable unexpectedly or as scheduled, a standby cluster will become the primary cluster. This can ensure system continuity and minimize downtime.
In addition, OceanBase Database allows you to back up data blocks and transaction logs for restore. This also ensures high availability.
Hybrid transactional and analytical processing (HTAP)
The innovative distributed computing engine of OceanBase Database allows multiple nodes to run online transaction processing (OLTP) and online analytical processing (OLAP) applications at the same time. This can maximize database utilization.
OceanBase Database uses one computing engine to support hybrid load balancing so that you can use only one system to resolve 80% of your needs. This makes OceanBase Database superior to distributed databases that use two types of computing engines or even two types of database systems to support both OLTP and OLAP.
Multitenancy
OceanBase Database uses a cloud database architecture and supports multitenancy in a cluster. You can deploy OceanBase Database in various types of clouds, such as the public cloud, private cloud, and hybrid cloud.
OceanBase Database achieves resource isolation by separating tenants. Database instances do not affect each other. OceanBase Database uses permission control to ensure the data security of all tenants. Combined with the extreme scalability, OceanBase Database can provide you with a secure and flexible database-as-a-service (DBaaS) service.
High compatibility
OceanBase Database supports Oracle and MySQL, which are the two most widely-used database ecosystems.
OceanBase Database supports all the syntax of MySQL 5.6. Business can be seamlessly switched between MySQL and OceanBase Database.
OceanBase Database also supports most Oracle syntax and almost all procedural language features. This helps implement automatic migration for most Oracle services after minor adjustments. OceanBase Database has landed in multiple finance enterprises in China and multiple internal systems of Alibaba Cloud.
Complete ownership of intellectual property rights
OceanBase Database is independently developed by Ant Group. It is not based on open-source databases such as MySQL or PostgreSQL. OceanBase Database is autonomous, controllable, and free from the technical limits of open-source databases.
High performance
OceanBase Database provides in-memory databases where data manipulation happens in the memory. OceanBase Database uses a storage engine that is based on log-structured merge-tree (LSM tree). This storage engine requires low hardware performance and well exceeds conventional relational databases in both read and write performance.
Security
The security features of OceanBase Database are designed based on the security requirement study on a large number of enterprises and various security standards. OceanBase Database can meet most security requirements of enterprises. It provides a complete permission and role system and supports features such as Secure Sockets Layer (SSL), transparent data encryption, auditing, label security, and the IP address allowlist. OceanBase Database is rated Level 3 in the evaluation of Classified Protection of Cybersecurity of China.
Core scenarios
Transparent partitioning for the payment system
Payment is the core business of Ant Group. The sharding solution was used in the beginning. Sharding provided this core service system with not only horizontal scalability but also the ability of canary upgrades, which significantly reduced system risks. However, after the rapid business expansion of Alipay, the capacity bottleneck of single nodes surfaced. In the sharding architecture, this bottleneck could be resolved only by breaking down data into M × N parts, which is time-consuming and has high technical risks.
This issue can be resolved by the partitioning feature of OceanBase Database, which brings horizontal scalability. OceanBase Database breaks down each of the N pieces of data into M data partitions. The partitioned tables can bypass the capacity limit of a single node. This way, OceanBase Database implements partitioning in the data layer without the need to modify the application system. This saves costs and lowers technical risks. At the same time, the partitioning feature of OceanBase Database can free you from considering the detailed data distribution. OceanBase Database also supports distributed transactions and JOIN operations across partitions, which allows transparent access to distributed databases.
Multidimensional query in membership systems
Many enterprises have membership systems that store important user information. User information is often queried from multiple dimensions. For example, you may want to query the detailed information about a user by using the ID or telephone number of the user, or group users by age. These multidimensional queries, especially those on non-partition key columns, are common in scenarios that involve massive amounts of distributed data. To handle these queries, you can create a shadow table that contains all the columns except for the partition key column. This approach uses no special database features. It is at best a workaround in the application layer. OceanBase Database can use the partitioning feature to break down data and distribute the data across multiple nodes in a cluster to meet some query requirements. Then, OceanBase Database uses strongly consistent global indexes to meet query and analysis requirements from other dimensions. This way, data sharding and horizontal scalability are implemented at the database level. This process is transparent to you.
Batch processing system
In many industries, batch processing systems need to process a large amount of data, such as the computation based on multiple joined large tables and data updates. In batch processing, a large amount of data is processed or updated at a time. Many large tables may need to be joined and complex queries may become inevitable. As a result, conventional centralized databases may frequently encounter single-point bottlenecks and incur high vertical scaling costs. The SQL engine of OceanBase Database has been continuously improved for more than 10 years. It can provide excellent SQL execution and distributed computing capabilities to support complex HTAP applications. As a cloud-native distributed database service, OceanBase Database does not have single-node bottlenecks and can help you save scaling costs.
Learn more about OceanBase Database
You can learn more about OceanBase Database by reading the following topics: