Each student will participate in the design and creation of a project that is intended to demonstrate their competence in building and analyzing a model of a computer or communication system, conducting analysis of that model, and reporting on the results discovered.

The scope of possible project topic areas is broad. Examples include

• A model of a cyber-physical system, including a model both of the physical part and the cyber part that controls it, to assess the consequences to the physical system of an attack on the cyber component, and to assess the impact of passive and active defenses against those attacks.
• A model of some portion of a CPU architecture (e.g. memory system, GPUs) with the objective of assessing some critical system attribute such as instruction throughput, or energy use, or resilience to manufacturing errors.
• A large model using Petri nets to describe the behavior of a distributed algorithm and determine the likelihood of the algorithm failing, or failing to achieve some performance goal.
• A model of a large scale data center, under AI-induced loads, that responds to unacceptable variations in the supply of commercial power by transitioning to the center's own power generation source, but at significant cost, or reduction of computational load, also at significant cost. The assessment might be to develop a strategy that minimizes response cost.
• A model of the switching network of a parallel supercomputer, that explores various strategies for routing, e.g., cut-through or store and forward, including models of applications running on the supercomputer that drive the communication traffic, with the objective of finding strategies that minimize overall execution time.
• A model of a communication network using satellites, that is under physical and/or cyber-attack. What is the impact on connectivity as a function of different state-aware algorithms that attempt to mitigate the attack.
• Other. Students encouraged to think up projects they will find interesting.

Multiple students may work on a project (but no more than four). Projects by individuals are permitted, the expected scope of any project will be in proportion to the group size.

Every project (and presentation) should cover the following points

• Domain overview, role of project topic in domain
• Research problems the study addressed, typically questions about some attributes of system behavior in defined contexts
• Description of the model being analyzed
• Description of the modeling tools and/or programs written to perform the experiments
• Description of the experimental design
• Presentation of results
• Conclusions and lessons learned.

Each project should be represented by a video, 10-20 minutes in length. Selected project teams may be asked to make presentation in the class.

Key Dates

Due Date	Deliverable
Oct 13	Last date for proposing (by email to Professor Nicol) a project team, and an intended project area
Nov 3	Draft description of model, and research questions to be addressed
Nov 17	Draft description of experimental design
Dec 8	10 - 20 minute video presentation

Text

No single textbook required. Each topic is extensively covered on-line, use provided lecture slides as guides. Lecture slides will be posted on Canvas at least one day before the class in which it is presented.

Prerequisites

ECE 313 or equivalent.

Overview

Development of analytic and simulation models of computer systems and application of such models to system analysis: latency, bandwidth, availability, reliability, security. The intent of the course is to prepare students to develop analyses of computer systems in application domains where they may work or perform research. The course focuses on methodologies, tools, and data structures required to perform such analyses. There is an emphasis on working through case studies that highlight different modeling and analysis techniques.

The course will emphasize hands-on in-class problem solving. Most weeks the first lecture will cover technical material, illustrating its use with workable examples. A week's second lecture will start with an in-class quiz on the previous lecture's material, and be followed by presentation of modeling problems the students are to work on in-class. Throughput the class the instructor and teaching assistant will be available for consultation and advice, the solutions will be required at the end of the class. Precise grading will not be applied to problem solutions, the point is to ensure students engage in the modeling process.

Owing to the challenges of performance assessment in the age of AI, all material used to assess students will be gathered in-person, in-class.

Grading Policy

Activity	Grade Composition
Weekly Quizzes	10%
Weekly Problem Sets	10%
Midterm 1	20%
Midterm 2	20%
Group Project	15%
Final Exam	25%

Course Rules

All exams will be in-class, closed-book, but will allow for two sheets of 8.5"x11" paper to contain notes available for consultation during the exam. For group projects students will be assigned to groups of up to 4, and the group will develop a project that models and analyzes some complex system using the tools and methodologies taught in the course, and make a class presentation on the project, its approach to modeling, and the results that can be concluded from the project's experiments. No written report is required, only the slides used in the presentation.