Numerous applications such as user activity history (e.g., web-site visit click and ad click) and financial transactions (e.g., stock trading) are write-heavy in nature. The backend storage systems used in these applications must support very high write throughput in addition to other important functionalities such as elastic scalability, multi-version data access, transactional semantics for bundled read and write operations, and fast recovery from machine failures.
The LogBase project aims to develop a scalable log-structured database system that meets the above requirements. Since log files constitute the only data repository in the system, every aspect of the system, for examples, novel indexing structures and query processing strategies, need to be tailored / redesigned for the new environment.
We are thankful to the ASTAR for supporting our research.
More projects in database research group in NUS can be found here!