Highlights from SDCC 2015 Database Practice Forum: Distributed Database Technologies and Real-World Implementations

SDCC, founded in 2007, has become a renowned technical conference brand; this article reviews the 2015 Database Practice Forum, featuring eight expert talks on distributed databases, performance optimization, and real‑world deployments.

1. Lu Yilei, Vice President of Technology at AdMaster – Distributed Database Challenges and Analysis He introduced isolation, the CAP theorem, and NoSQL concepts, compared OLTP and OLAP from user‑centric and system perspectives, discussed SSD‑based performance boosts, distributed transaction support, and real‑time query engines such as Elasticsearch, Pinot, and Druid.

2. Wang Jingyu (alias Shen Xun), Senior Technical Expert at Alibaba – Alibaba Distributed Database DRDS in Double‑11 He traced the evolution of database models, highlighted the need for horizontal scaling in the mobile/Internet era, and described how DRDS powered Alibaba’s Double‑11 sales, scaling TPS from 5,000 to several million with multi‑data‑center deployments.

3. Liu Qi, Founder of TiDB – TiDB, HBase Distributed Transactions and SQL Implementation He explained TiDB’s architecture for high‑concurrency, large‑scale distributed systems, its MySQL‑compatible protocol, and detailed the implementation of TiDB’s distributed transaction engine and how distributed transactions are realized on HBase.

4. Ma Ruyue, Director of Architecture at Baidu Big Data Department – Palo Analytic Database in Baidu He presented Palo’s role in statistical analysis and reporting, showing how Baidu reduced machine count from 220 to 58 and cut average query latency from 60 ms to 30 ms, and described its multi‑model architecture and key technologies.

5. Liu Ruimin, Chairman & CTO of Baidu Data – RapidsDB Massive Parallel SQL for Deep Learning He described RapidsDB’s physical and logical architecture, its distributed query engine DQS, and how users write plain SQL while the system handles memory‑distributed processing, query rewriting, and MPP optimization for massive parallel workloads.

6. Sun Guorong, Technical Director of GBase 8t at Nanda General – GBase 8t High‑Availability Solution He explained GBase 8t’s database‑level HA capabilities, including automatic failover within seconds, to address challenges such as node crashes, disk failures, and network interruptions.

7. Pan Anqun, Senior Software Engineer at Tencent – Tencent Financial‑Grade Database TDSQL Analysis He detailed TDSQL’s MySQL‑based core, its use in products like Mi Master, Webank, and Tencent Cloud, and its requirements for high consistency, zero data loss, 24/7 automatic disaster recovery, high scalability, and support for billions of daily transactions.

8. Wang Tao, Co‑Founder of SequoiaDB – SequoiaDB and Spark in the Financial Industry He introduced SequoiaDB as a NoSQL solution, its collection‑partition mechanism for time‑series data, its ability to avoid index‑tree bloat, and its integration with Spark for interactive analytics as well as compatibility with Hadoop ecosystems.

Follow the ArchDigest public account for daily high‑quality articles on architecture, big data, machine learning, and other hot technology topics.