|
|
MIT Electrical Engineering and Computer Science
Spring 2002 Catalogue Supplement |
R2-5, Room 1-379
Dr. Michael Stonebraker, NE43-511, 3-0130
Prereq.: Undergraduate course in database systems or permission of instructor
3-0-9
Catalog description: This course will cover the important subject matter relating to the architecture of database systems. We treat data models and query languages, including hierarchical, network, relational and the recent W3C proposals in this area. Also discussed are the implementation of database systems, including query processing, view management, transaction processing, access methods, buffering and storage management. Wrinkles appropriate to read-only data (warehouses), dedicated hardware (database machines) and distributed data will also be covered. We then discuss the impact the web is having on DBMS implementations, and finish with issues from several niche markets, including geographic information, real-time sensor data, time series data, and media asset management
Expanded description: This course presents the important ideas in the design and implementation of database systems. First, the various data models and query languages that have been proposed will be covered. The emphasis will be on why the candidates have succeeded or failed and what there is to learn from the past. Then, we discuss the software components that comprise a DBMS. Important topics are query optimization, transaction support, access methods and storage management. In each case, we present the state of the art as well as speculative ideas. Many organizations move historical data to a centralized data warehouse for business analysis purposes. The special needs of warehouses will be addressed, as well as the issues surrounding data that spans multiple sites in a computer network. To date, there have been no successful database machines that use customized hardware. We will treat the various attempts, and indicate why we do not expect to see successful products in this area in the future. It is often said that "the web changes everything", and we treat the impact that the web is having on DBMS architecture. We also discuss the possibilities for caching structured data at the edges of the web. We finish the course with issues that arise in several niche markets, and discuss the special needs of geographic data, time series data (such as stock market ticker data), real-time sensor data (such as tracking assets like cell phones), and the storage of large amounts of video.
Text: Readings in Data Base Systems, 3rd edition by Joey Hellerstein and Michael Stonebraker, plus other papers from the literature
Workload: There will be a midterm, a final exam, two short research papers and a term project.