15-648: Big Data Systems Studio (Spring 2014)

Overview

Advancing the state of the art in large scale systems is challenging for many reasons including algorithm efficiency, weak consistency, failure tolerance, etc. However, as observed repeatedly among the Googles, Facebooks and Yahoo!'s, all the rest of these hardly matter if the developer is not experienced with large systems, if the logistics of large scale systems are not easily mastered, and if the large collection of core systems is unfamiliar. Without good hands on experience with the current systems and their tools, theoretical knowledge leads to theoretical success, and too often little more. This course provides hands on experience and a sampling of the latest systems research.

This course is a project course intended primarily for the Big Data Systems track of the Computational Data Science (formerly Very Large Informations Systems) program, Systems track. Others need instructor approval to register. The course has two components: readings of current operating and distributed systems research, and a novel big data systems project.

The readings will be drawn from the most recent year or two from the following systems conferences: Symposium on Operating Systems, Symposium on Cloud Computing, Operating Systems Design and Implementation, Networked Systems Design and Implementation. For one paper a week each student submits a one page review composed of citation, a paragraph or two exposition of the paper's content, three strengths, explained, and three weaknesses, explained. Send reports under the title "648 Review of paper-title" to garth+15648 at cs dot cmu dot edu.

Projects will be individually done, starting with an instructor approved proposal and ending with a presentation and final report (about the length and format of a research paper). See the Project page for more information.

• Lecture time: Fridays 13:30 - 16:20
• Location: Wean 4623
• Units: 12
• Prerequistes: Students must be enrolled in the Computation Data Science Systems Track, or have instructor permission.
• Emailing course staff: 15648-staff at mailman dot srv dot cs dot cmu dot edu
• Web site: www.cs.cmu.edu/~15648