OBKV-Table is a distributed table storage system that provides better throughput and lower latency than SQL through simple interfaces. This topic describes OBKV-Table in detail.
Application scenarios of OBKV-Table
OBKV-Table is not just a simple key-value database. It also provides secondary indexes, data types, conditional filtering, and aggregation, which meet the requirements for common relational database capabilities in simple OLTP scenarios. Typical characteristics are as follows:
- The business scenario involves only single-table operations.
- The business scenario involves only single-partition operations, including two cases: all data operations use the primary key, or all data operations use a primary key prefix that is also the partitioning key.
- The business scenario requires extremely stable latency or higher throughput than SQL.
Scenarios where OBKV-Table is not suitable
We do not recommend OBKV-Table if any of the following conditions applies to your business scenario:
- Presence of multi-table join queries.
- A large number of distributed transactions.
- Complex query operations.
Advantages of OBKV-Table
Performance:
- It eliminates cross-node transactions, cross-node data interaction, and complex data computation, ensuring the shortest data path and better P99 tail latency.
- It provides simple external interfaces that directly use native storage capabilities, delivering performance more than 30% better than that of OceanBase AI Database SQL.
Throughput and latency:
- All operations are deterministic, without complex issues such as non-optimal SQL plan execution or remote plans, which ensures deterministic throughput and latency.