ApsaraDB for OceanBase

A financial-grade distributed relational database that features high stability, high scalability, and high performance.

ApsaraDB for OceanBase is a relational database service developed by Alibaba Group for high-throughput, high-concurrency, and high-availability scenarios. The maximum processing capacity is 61 million times per second. The maximum data volume in a single cluster is more than 3 PB, and the maximum number of table rows is more than one trillion. ApsaraDB for OceanBase uses the Paxos protocol and maintains multiple data replicas. You can build a financial-grade database by deploying PC servers in a distributed architecture. ApsaraDB for OceanBase allows you to achieve reliability and data consistency at one third of the cost of other commercial databases.

Benefits

Outstanding Performance: Processing Capacity of 61 Million Times Per Second
ApsaraDB for OceanBase supports rapid business scaling and achieves high performance by using the quasi-memory processing architecture. The maximum data volume in a single cluster is more than 3 PB, and the maximum number of table rows is more than one trillion.
Continuous Availability: Leader and Practitioner of the Five Data Centers in Three Regions Solution
ApsaraDB for OceanBase uses the Paxos protocol and maintains multiple data replicas. ApsaraDB for OceanBase supports automatic switchover between common PC servers, where no data record is lost (RPO = 0, RTO < 30 seconds). This avoids single point of failures (SPOFs) and zone-related risks.
High Compatibility: Compatible with MySQL and Oracle
ApsaraDB for OceanBase is compatible with MySQL and Oracle. You can use ApsaraDB for OceanBase after making minor changes to existing code and applications.
Time-Tested Capabilities: Best Practices of Alipay and MYbank
ApsaraDB for OceanBase has been running in all core scenarios of Ant Financial for many years. The database is proved to be reliable during the Double 11 in the last ten years.
Unlimited Scalability
Distributed transactions and global timestamps ensure that the computing and storage capacity can be scaled out without code modifications. This eliminates the need to worry about table sharding and database sharding.
Proprietary Technology: Relational Database Developed by Alibaba Group and Ant Financial
ApsaraDB for OceanBase is different from open-source databases such as MySQL and PostgreSQL. Alibaba Group and Ant Financial have proprietary rights to all code of ApsaraDB for OceanBase. The troubleshooting and architecture optimization of all product modules can be achieved based on code.

Features

Architectures for Various Availability Requirements

ApsaraDB for OceanBase can be deployed across zones and regions. If a fault occurs, your workloads are switched to a healthy zone based on the specified priority. You can select zones based on your business requirements.

Zone-Disaster Recovery

You can deploy ApsaraDB for OceanBase across three zones in a single region. In this scenario, you can achieve a Recovery Point Objective (RPO) of zero and a Recovery Time Objective (RTO) of less than 30 seconds. Automatic failover is triggered when a fault occurs. This deployment solution helps you ensure service continuity when a hardware fault or zone disaster occurs.

Geo-Disaster Recovery

You can deploy ApsaraDB for OceanBase across five zones in three regions. In this scenario, you can achieve an RPO of zero and an RTO of less than 30 seconds. Automatic failover is triggered when a fault occurs. This deployment solution helps you ensure service continuity when a hardware fault, zone disaster, or geo-disaster occurs.

Multitenancy to Meet Business Requirements at Minimal Cost

ApsaraDB for OceanBase allows you to select the database engine for a cluster.

Oracle

ApsaraDB for OceanBase is compatible with the Oracle syntax and stored procedures. Oracle-based applications can be migrated seamlessly or with minimal modifications.

MySQL

ApsaraDB for OceanBase is compatible with the MySQL 5.6 syntax and client. You can use ApsaraDB for OceanBase the same way as you use MySQL. ApsaraDB for OceanBase provides financial-grade database capabilities such as scalability and high availability.

Financial-grade Features

ApsaraDB for OceanBase provides financial-grade features such as recycle bin and flashback query to ensure service availability and fault tolerance.

Flashback Query

ApsaraDB for OceanBase provides the flashback query feature for DML-related misoperations. This feature allows you to query previous data values or restore a table or row to a previous time point.

Recycle bin

ApsaraDB for OceanBase provides the recycle bin feature for DDL-related misoperations. This feature allows you to restore data in case you perform the DROP TABLE or DROP DATABASE operation.

Execution Plan Evolution

Only verified execution plans can be enabled. When a new plan is available for the target SQL statement due to various reasons such as changes in statistics, the new plan is not immediately enabled. The new plan must be verified by a small volume of data inside the OBServer and must have better performance than previous plans before it can be enabled.

Customer Success Stories

Related Resources

High-performance database

Outstanding performance during TPC-C benchmark test

All database kernel developers want to take on the challenge of the TPC-C benchmark. However, it is never easy to address all of the challenges that are presented in complex benchmarks like TPC-C. This also happened to be the very first time that ApsaraDB for OceanBase took on the TPC-C benchmark. To beat the TPC-C benchmark test, the research and development team of ApsaraDB for OceanBase prepared for more than a year.

High-availability database

High availability and disaster recovery

The primary requirement for high availability of databases is automatic failover with zero data loss. If automatic failover is enabled, services can be recovered in the shortest time. Recovery Time Objective (RTO) is used to measure the recovery time. Recovery Point Objective (RPO) is the metric for data loss after service recovery. RTO and RPO are important metrics to evaluate a high availability solution.

High-throughput database

Alipay technology during the Double 11

During the Double 11 in 2019, CNY 10 billion (USD 1.4 billion) in transactions were completed in 1 minute and 36 seconds, CNY 30 billion (USD 4.2 billion) in 5 minutes and 25 seconds, and CNY 50 billion (USD 7.1 billion) in less than 13 minutes. The Double 11 has allowed Alibaba Group to lead the way in terms of research and development. Alibaba Group has advanced the level of computing capabilities and related technologies by approximately 20 years.

Clusters

Create a cluster

A cluster is a basic unit for purchasing cloud products.

Tenant

Create a tenant

Tenants are constituent parts of a cluster. You can create tenants to make full use of the system resources in the cluster.

Database

Create a database

You can create different databases for independent tenants to store data.

Scenarios

  • High Availability and Disaster Recovery
  • Scale-out
  • Global Consistency
Data Migration

Data Migration

Data Migration to a Distributed Database

ApsaraDB for OceanBase provides an end-to-end data migration service for enterprises to migrate applications and data to ApsaraDB for OceanBase. This solution aims to bring the benefits of distributed databases to more enterprises.

  • Assessment of Application Changes

    The data migration service of ApsaraDB for OceanBase helps you assess the impact of migration on your databases. For example, you can identify incompatible SQL statements and schemas to prevent risks of project delays.

  • Innovation

    Changing the database of a business system poses significant challenges and risks to data accuracy, performance, and stability. The data migration service of ApsaraDB for OceanBase is integrated with a multi-level data verification scheme that helps you verify the performance and compatibility of databases. You can detect data exceptions within seconds. ApsaraDB for OceanBase also provides traffic capture and playback capabilities.

High Availability and Disaster Recovery

High Availability and Disaster Recovery

Unattended and Self-Adaptive Disaster Recovery

ApsaraDB for OceanBase provides a customizable solution for high availability and disaster recovery. You can deploy ApsaraDB for OceanBase based on the application scenarios and your business requirements for zones and regions. You can implement financial-grade high availability by using basic hardware.

  • Flexible High Availability

    ApsaraDB for OceanBase can be deployed across zones and regions. It allows you to customize the disaster recovery capability based on your business scenarios. ApsaraDB for OceanBase uses the Paxos protocol to ensure the consistency of data replicas that are deployed across different regions or zones. If a zone or region fails, ApsaraDB for OceanBase switches your workloads to a healthy node where a data replica resides.

  • High Availability Across Regions and Unit-based Redundancy

    If you deploy ApsaraDB for OceanBase across regions, you can deploy the master nodes of different tenants in different regions. With OBProxy, you can modularize your business systems and realize high availability across regions. Among users of ApsaraDB for OceanBase, MYbank and Paytm adopt similar architectures. Module-based distribution and redundancy provide support for capacity scaling and region-level disaster recovery.

  • Asynchronous Replication

    ApsaraDB for OceanBase asynchronously replicates data from the active cluster to the standby cluster. The active and standby clusters are independent of each other. This prevents failures from being transferred to the standby cluster when the active cluster encounters an error.

Scale-out

Scale-out

Similar Experience to Using Standalone Databases

ApsaraDB for OceanBase is a native distributed relational database that can be used in the same way as a standalone relational database. You can improve the computing and storage capacity by increasing the number of nodes. In a distributed system or sharded architecture, global indexes and global consistency are usually compromised. Therefore, it requires extra costs to address these needs. To solve this issue, ApsaraDB for OceanBase provides capabilities such as snapshot isolation levels and multi-version concurrency control (MVCC) by using global timestamps and global indexes.

  • Distributed Transactions

    In ApsaraDB for OceanBase, a table is split into many small partitions and distributed across multiple nodes for storage based on a specific policy. All transactions are executed as distributed transactions. Global timestamps are used to ensure that the distributed transactions in an ApsaraDB for OceanBase cluster meet the requirements of atomicity, consistency, isolation, and durability (ACID) and transaction isolation levels.

  • Global Indexes

    ApsaraDB for OceanBase supports partial and global indexes within a partition. Partial indexes have lower maintenance costs and higher execution efficiency. Global indexes can help avoid partition scans and table scans that are required in traditional sharded architectures. Global uniqueness constraints can help avoid complex architectures introduced by various uniqueness requirements. ApsaraDB for OceanBase optimizes the performance of global indexes by using the two-phase commit mechanism and non-covered queries. The optimization ensures that global indexes can meet the requirements of most core scenarios.

Global Consistency

Global Consistency

Transaction Integrity Across Databases

ApsaraDB for OceanBase provides a global consistency solution that does not rely on specific hardware. ApsaraDB for OceanBase provides capabilities such as snapshot isolation levels and multi-version concurrency control (MVCC) to ensure global consistency of snapshot versions.

Traditional architectures for table sharding and database sharding do not support global and unified snapshot management.

  • Transaction Integrity

    ApsaraDB for OceanBase uses Global Timestamp Service (GTS) to ensure consistency of snapshot versions. All transactions obtain snapshot versions from GTS regardless of the source servers. The version numbers of snapshots increase over time. The timestamps of all snapshots are consistent with the real world chronological order. ApsaraDB for OceanBase assigns global timestamps by tenant.

Get Started with ApsaraDB for OceanBase

Learn more about ApsaraDB for OceanBase

FAQ > Console >