When selecting the underlying database for its custom AI assistant, Quwan Technology compared multiple vector databases and chose OceanBase. This article shares its thinking and experience in building a three-in-one database foundation for scalar, vector, and full-text search that meets the multiple needs of AI applications.
Qifu Technology adopted OceanBase to replace part of its MySQL and TiDB usage, cutting code-refactoring costs by 90%, halving storage costs, reducing servers by 30+, and shortening task execution time by 40%. This article shares the database pain points across multiple scenarios and Qifu's hands-on experience deploying OceanBase.
Exploring a new form of combining AI with database products. This article introduces how the obloader agent simplifies the use of data-loading tools through AIChat, and how the OceanBase Agent solves the multi-database management pain points of general-purpose MCP clients, enabling natural-language interaction to manage database instances.
NetEase Cloud Music shares how it migrated its PB-scale sharded DDB database to the native distributed database OceanBase. Built on its in-house CDC service NDC, the migration was transparent to the business and lost zero data, with forward synchronization reaching GB/s throughput and storage shrinking by at least 1/4.
The official guide to using the OceanBase vector database: a detailed walkthrough of vector index types such as HNSW, HNSW_SQ, HNSW_BQ, IVF, and IVF_PQ, covering type selection, parameter tuning, memory estimation, and partitioned-table practice, with performance optimization recommendations spanning data volumes from millions to hundreds of millions.
This article shares the complete process of deploying a standalone OceanBase Community Edition database via OBD's GUI in a WSL Ubuntu 22.04 virtual machine on a Windows laptop. It covers environment preparation, how to install RPM packages under Ubuntu, cluster and component configuration, the principles behind ob_configserver, and the OB Dashboard operations experience.
This article systematically reviews the fault classification and root causes of OceanBase, proposes a five-step diagnosis and tuning process—problem identification, data collection, problem localization, solution formulation, and validation—and introduces key tools and methods such as internal views, log analysis, OCP, OAS, and obdiag.
When an intelligent robot cannot understand minor languages such as Thai, developers can use OceanBase's plugin mechanism to respond quickly to requirements. This article explains the plugin architecture's low coupling, openness, and self-controllability, as well as how to develop and use tokenizer plugins and external table plugins—helping you bypass the long kernel development cycle and quickly extend database capabilities.
A card game company built an intelligent customer service and UGC community recommendation system on OceanBase's vector database. Through HNSW indexing, supplementary keyword retrieval, and intelligent reranking, it achieved millisecond-level retrieval, a 300% boost in customer service efficiency, a 75% reduction in recommendation latency, and a 75% reduction in operational complexity.
NetEase Personal Mail upgraded its sharded MySQL architecture to OceanBase. Through technology selection comparison, unique-key governance, partition and index optimization, and OBKV-Redis practice, it achieved 72% storage cost savings, a significant QPS increase, and a real-time analytics upgrade.