Fiches de cours

Systems for data science

CS-449

Lecturer(s) :

Kermarrec Anne-Marie

Language:

English

Summary

The course covers fundamental principles for understanding and building systems for managing and analyzing large amounts of data.

Content

Big data systems design and implementation :

• Distributed systems for data science
  
• Data management : locality, accesses, partioning, replication
  
• Distributed Machine Learning Systems : federated learning/parameter server/decentralized learning
  
• Massively parallel processing operations

Large-scale storage systems :

• Data structures : File systems, Key-value stores, DBMS
  
• Concurrent access to data
  
• Consistency models. The CAP theorem. NoSQL and NewSQL systems
• Transactions
  

Large-scale processing :

• Parallel processing
  
• Streaming Processing
  
• Online Processing
  
• 'Graph Processing
  

Keywords

Distributed systems, Parallel programming, Large-scale storage systems, Large-scale data management

Learning Prerequisites

Required courses

CS-322: Introduction to database systems

Recommended courses

CS-323: Introduction to operating systems

CS-206 Parallelism and concurrency

Important concepts to start the course

• Algorithms and data structures ' sorting algorithms, balanced trees, graph traversals.
• The Scala programming language will be used throughout the course. Programming experience in this language is strongly recommended.
• Basic knowledge or computer networking and distributed systems

 

Learning Outcomes

By the end of the course, the student must be able to:

• Choose systems parameters, data layouts, query plans, and application designs for database systems and applications.
• Develop data-parallel analytics programs that make use of modern clusters and cloud offerings to scale up to very large workloads.
• Analyze the trade-offs between various approaches to large-scala data management and analytics, depending on efficiency, scalability, and latency needs
• Choose the most appropriate existing systems architecture and technology for a task

Teaching methods

Ex cathedra; including exercises in class, practice with pen and paper or with a computer, and a project

Expected student activities

During the semester, the students are expected to:

• attend the lectures in order to ask questions and interact with the professor,
• attend the exercises session to solve and discuss exercises,
• solve practical homeworks and/or finish a project during the semester,
• take a midterm
• take a final exam

Assessment methods

Homeworks, written examinations, project. Continuous control

Supervision

Office hours	Yes
Assistants	Yes
Forum	Yes
Others	Office ours by appointment

Resources

Bibliography

Relevant resources (textbook chapters, articles, and videos) posted on moodle page.