This is a first-year graduate class in distributed systems. Distributed systems are a powerful tool that enables scaling storage and computation power at relatively cheap costs. However, distribution brings in many complexities unseen in single computer systems. In this course, students will learn various such complexities and current solutions, both from industry and academia. This course requires reading 1 or 2 papers per week, four lab projects, and an (optional) final project. Each class will have a mix of lecture and discussion.
Students will learn
How large-scale systems are structured
Different replication techniques and their tradeoffs
How to build a fault-tolerant systems
Data consistency challenges and a few remedies
How to speed up big data processing with data parallelism
Emerging distributed programming model and framework
How distributed systems may improve energy and cost efficiency
Professor: Seo Jin Park
TA: Yizhuo Liang, Junzhou He
Office hours: announced on Piazza
Questions: piazza (To ensure response, do not send emails to instructors directly.)
Time: Tue/Thurs, 2:00 - 3:50pm
Location: THH 102
Exam (45%), programming lab assignments (35%), homework (10%), in-class participation (10%)
(Optional) Final project may replace the midterm or lab3.
There are no textbooks for this class. All materials will be covered in lecture or literatures assigned for reading.