Systems Seminar

URL study guide

https://studiegids.vu.nl/en/courses/2024-2025/XM_0122

Course Objective

The key objective of this seminar is to introduce students to designs, details, and challenges associated with building modern computer systems (CompSys) from a single machine to distributed services. As part of this exercise, students will read, review, present, and discuss recently published cutting-edge research on the design and implementation of computer and networked systems. The course takes the form of an interactive seminar (presentations and discussions) mostly driven by the students. Upon completion of this course, students will be able to:Reason about various design choices, systems building trade-offs, and ongoing and emerging research challenges for a design based on the scale, deployment goals, and capabilities provided. [Applying knowledge and understanding] [Making judgements]Understand recent advancements in systems (leveraging modern hardware
- SmartNICs, fast storage, FGPAs, programmable switches) and cutting-edge techniques to run these systems safely and efficiently. [Knowledge and understanding]Read papers, write critical (technical) reviews (style, tone, details), compare details across other papers, understand how to evaluate a paper, and develop and learn about research methodology skills. [Knowledge and understanding] [Applying knowledge and understanding] [Communication]Learn how to present and lead discussions on state-of-the-art literature in an interactive setting with your peers [Communication] [Lifelong learning skills]

Course Content

The course is largely student-driven. Each week we will have paper presentations and discussion sessions where we will discuss recently published, high-impact papers in the literature. We will provide a curated list of recently published papers in the first week of the class to cover these topics. The acquisition of this knowledge is important to understand how the fields of computer systems and networks are emerging, and how future cloud/distributed systems in practice will be shaped. In each class session there will be a presentation(s) from students, and students will lead the discussion on the topic. Outside the class, the students need to read papers and write detailed technical reviews.

Teaching Methods

In each lecture session we will discuss a few closely related papers in detail. Every student is expected to read all papers and send weekly reviews before the week starts. A sample format for the review and previous year's review samples will be provided. There will be no deadline extensions and we will not allow late submissions. We will, however, provide additional papers to review to make up for the lost marks. In each lecture there will be presentations (15-20mins) from students, and then an open, conference program committee (PC) style discussion led by students regarding the (i) content and contributions; (ii) strengths; and (iii) weaknesses (and what could be improved); (iv) how papers which are discussed are related to each other; (v) various other issues raised in your submitted reviews; and (vi) overall broader research direction to which the papers belong. For this seminar we expect the students to actively participate by means of presentations and discussions.

Method of Assessment

Students are awarded points every week on the quality of the review (technical correctness, understanding of the paper, details covered, tone and style), and the presentations (style, delivery, visuals, and audience engagement). The final grade is a weighted average of points as: Reviews (60%) and presentations (40%). You must score at least 5.5 in each of the two components (i.e., reviews and presentations). The final grade is calculated as a weighted sum of points obtained for the reviews and presentation, as shown above. There are no resits available for this course. [Fraud] The course is made fraud resistant by having selective and specialized reviews to assess the knowledge of the students which are graded individually.

Literature

A list of recently published and classic papers (mandatory and background readings) reflecting recent developments. A sample format for the review and previous year's review samples will be provided. We will also introduce advanced topics, reproducibility research methodology, and tips for reviewing and presenting works.

Target Audience

MSc Computer SecurityMSc Computer Science

Recommended background knowledge

This course builds on a wide variety of computer systems and networking knowledge and concepts such as design of operating systems, networking and storage stacks, integration of accelerators, memory virtualization and management, network architecture, Internet-scale computing, process management, failure and fault tolerance in distributed systems, computer architecture (interrupts, CPU caches, multi-core systems, I/O buses), etc. Due to the nature of the course, there will be limited opportunity to learn these concepts in the class. In case you lack necessary background knowledge, ask us early to provide background reading material that you should cover by yourself.