SIGMOD 2018 Conference Report and Paper Highlights by Ant Group OceanBase

SIGMOD 2018 was held for the first time in the southern United States, in Houston, attracting about 800 attendees from around the world to discuss the latest advances in database research and practice.

The report is written by Rizhao, a researcher from Ant Group's OceanBase team and one of the core contributors to OceanBase versions 0.5/1.0/2.0, who offers a professional and up‑to‑date coverage of the conference.

The conference featured two keynote talks—Eric Brewer’s "Kubernetes and the New Cloud" and Pedro Domingos’s "Machine Learning for Data Management: Problems and Solutions"—and a strong industry presence with sessions from Amazon, Microsoft, Alibaba, and others, highlighting the growing importance of industry contributions.

The best paper award went to CMU’s "SuRF: Practical Range Query Filtering with Fast Succinct Tries", which introduces the Fast Succinct Tries (FST) data structure for efficient range filtering, improving on traditional Bloom filter approaches.

Five major research trends were identified from the program: cloud databases, emerging hardware (SSD, NVM, RDMA, GPU, FPGA), autonomous databases, AI‑augmented databases, and graph databases.

Eight industrial papers were highlighted: 1. Amazon Aurora – storage‑compute separation and log‑shifting architecture; 2. Microsoft SCOPE – online materialized view framework CloudView for reducing duplicate computation; 3. Hybrid B+‑tree/Columnstore – performance comparison and dynamic index selection; 4. P‑Store – predictive provisioning for elastic databases; 5. Pinot – real‑time OLAP system for massive user bases; 6. Vertica EonMode – cloud‑native columnar database with storage‑compute separation; 7. Azure SQL survivability – machine‑learning model to predict database lifespan; 8. Hyper‑Q – middleware for seamless migration of SQL workloads to the cloud.

The article concludes with an invitation to follow the OceanBase official WeChat account, join the technical community, and stay updated on future SIGMOD insights.