CS 4320 (and 5320): Introduction to Database Systems

CS4320/CS5320 is an introduction to relational database systems, NoSQL and NewSQL systems, and other tools for large-scale data analysis. Topics covered include the relational model, SQL, query processing and optimization, transactions, recovery, NoSQL and NewSQL systems, database design, as well as systems for graph, stream, and spatial data processing.

All assignments will be distributed over the Course Management System (CMS). We will be using Ed as a forum for student discussion, for announcements, and to distribute material such as lecture notes. Therefore, this Web site is deliberately fairly minimal.

Course registration link for the Ed discussion forum: here.

Staff

Here are the contact details for the course staff:

• Instructor: Immanuel Trummer (it224@cornell.edu)

Lectures

The course meets three times a week, MWF 2:40pm - 3:30pm in Olin Hall 155. Video recordings and slides from last year are available on www.databaselecture.com.

Workload

• The course has five homeworks which can be solved in groups of up to three students.
• The course includes one final and one preliminary exam.
• The CS 5320 version of the course includes one course project.

Course Project

The course project is only required when taking the CS 5320 version of the course (not for the CS 4320 version). The goal of the course project is to analyze a recent, scientific paper from the database domain. One of the following papers can be selected for the course project (selecting a different paper requires explicit permission by the course instructor):

• Efficient join algorithms for large database tables in a multi-gpu environment.
• Jointly optimizing preprocessing and inference for DNN-based visual analytics.
• Zen: a high-throughput log-free OLTP engine for non-volatile main memory.
• Adaptive code generation for data-intensive analytics.
• NOAH: interactive spreadsheet exploration with dynamic hierarchical overviews.

Projects are solved by groups of up to six students. Each team submits a one-page document about the paper until the end of the semester. This document contains the following parts:

• A summary of paper content, discussing what is proposed, what the main ideas are, and how that differs from prior work.
• A short discussion (i.e., few sentences) of how the paper content relates to other course material.
• Three advantages of the solution proposed in the paper over prior work.
• Three propositions of how the paper solution could be improved or extended.

To achieve full points, the submitted document must provide a detailed and accurate summary, as well as reasonable and well justified advantages and proposals for extensions.

Grading

The grade is based on a weighted average of the different course components:

• Each of the five homework is counted with weight 1.
• The preliminary exam is counted with weight 2.
• The final exam is counted with weight 3.
• For the CS 5320 version, the course project is counted with weight 2.

Syllabus

The detailed course syllabus is available here. Slides and recordings of past semesters are available on www.databaselecture.com.

Office Hours

Changes to regular office hours will be posted on Ed. Please make sure to check on Ed before visiting an office hour!

Additional Material

About the first two thirds of the course are based on the textbook "Database Management Systems" by Raghu Ramakrishnan and Johannes Gehrke. The last third of the course discusses recently proposed systems, more details about them can be found in the corresponding research papers.