Read:
Stonebraker, Michael, et al. C-Store: A Column-oriented DBMS. Proceedings of the 31st VLDB Conference, 2005. (PDF)
This paper discusses the C-Store system, which is a read-optimized database system with a novel physical arrangement of data on disk, as well as an unusual approach to recovery and query processing
As you read the paper, consider the following questions:
- Why does the C-Store idea of arranging data as columns improve query performance?
- How does C-Store propose to do recovery? Is there an advantage to this approach?
- How does C-Store isolate transactions? Do you think this is a good idea?