Intelligent Operating Systems

Basic Course Information

• Course Name: 지능형운영체제 (Intelligent Operating Systems)
• English Name: Intelligent Operating Systems
• Course Type: Major Elective (Track Required Recommended)
• Credits/Hours: 3 credits / 3 hours per week (Recommended: 2 lecture + 1 lab)
• Recommended Year: 2-3

Course Overview

This course covers the core principles of operating systems (processes/threads, scheduling, memory management, file systems, synchronization, security) while also addressing resource management perspectives required by modern AI workloads (accelerator utilization, container-based deployment, service latency requirements). While maintaining the identity of traditional operating systems, it connects to performance bottlenecks, isolation, and observability frequently encountered in actual AI system operations.

Educational Objectives

1. Understand core operating system abstractions and system call-based execution models.
2. Analyze concurrency and resource contention problems and apply appropriate synchronization techniques.
3. Explain virtual memory/file systems/I/O paths and analyze performance bottlenecks.
4. Understand container isolation and resource limitation concepts and verify them through experiments.

Learning Outcomes

• Explain the operating principles of processes/threads/scheduling.
• Reproduce deadlocks and race conditions and propose resolution strategies.
• Measure memory/file/I/O bottlenecks and suggest improvement directions.
• Write experimental reports utilizing container-based resource limitations and observability tools.

Prerequisites

• Required: Data Structures, Computer Architecture or Computer Systems Fundamentals
• Recommended: C or System Programming Fundamentals, Linux Usage Experience

Main Content (Modules)

• Processes/threads, context switching, scheduling fundamentals
• Synchronization (mutex/semaphore/monitor), deadlock
• Virtual memory, paging, memory hierarchy
• File systems, I/O, cache/buffer
• Virtualization overview, container (namespace/CGroups) resource limitations
• Performance measurement and observability (logs/metrics) fundamentals from an AI workload perspective

Practice and Assignment Examples

• Educational kernel assignments (based on xv6/Nachos, etc.) or system call/scheduling experiments
• Container resource limitation experiments (CPU/MEM) and performance comparison reports
• Simple inference server load experiments and bottleneck analysis (profiling/metrics)

Evaluation Method (Example)

• Theory Exam: 30%
• Lab Assignments: 45%
• Project: 15%
• Participation: 10%