Where: ECS 210I (CSEE conference room)
When: MW 4:00 - 5:15 PM
Who: Prof. Ethan Miller
The reading list is available online; it includes a schedule for the class. Please check it regularly to keep up with changes in paper assignments and checkpoints for your projects.
Computers have made great strides in processing power over the last decade. High-speed uniprocessors, and now parallel processors, allow users to consume and generate ever-greater quantities of data. However, sheer number crunching ability is only part of the picture; where does all of this data come from, and where does it go?
This seminar will address these questions by examining file systems and storage mechanisms used to manage the terabytes of data stored at scientific computing centers. We will read 1 - 2 papers per class on topics including disk arrays, parallel and high-performance file systems, tertiary storage devices, and archival file systems. We will also cover some scientific applications and discuss the demands they place on the storage system.
We will read between two and four papers each week; the reading list is available online. Each paper will be assigned to a student in the class, who will be responsible for leading the class discussion on it. In addition, all students will be expected to summarize (in one or two pages) the material covered each week. There will be a term project, which can be anything from a more extensive literature survey to (preferably) a research project exploring topics discussed in class.
Students taking this class should be somewhat familiar with both operating systems and computer architecture. Familiarity with some domain of scientific computing would also be helpful, though it isn't required.