CS7333 Communication Network Analysis and System

Instructor: Prof. Zhenzhe Zheng, zhengzhenzhe@sjtu.edu.cn

Office Hours: 3:00-4:00 Tue, 3-509 SEIEE.

Website: https:zhengzhenzhe220.github.io/CS7333index.html

Course Meeting Times: 18:00-20:20 Monday (Weeks 1-16) (陈瑞球楼 430).

Prerequisites: Basic knowledge of optimization, probability and linear systems or consent of instructor.

Credit: 3

Description

The goal of this class is to develop understanding of some fundamental techniques used to model and analyze communication networks: Performance analysis and design of multiple-user communication systems; emphasis on rigorous formulation and analytical and computational methods, includes queuing networks, decentralized minimum delay routing, and dynamic network flow control. These analytical tools are used to analyze the performance of various networks. More importantly, understanding this material can help one to develop intuition about some of the important issues in networking and provide the background needed to do research in this field.

Intended Audience

The course is geared towards computer science students who would like to have an in-depth understanding of modern communication networks for computer and computing, know more about fundamental issues, engineering tradeoffs.

Grading

• Homework: 50%

• Final Project: 30%

• Attendance and Scribing：20%

Topics (as time permits):

Analysis Part:

• End-to-end network architecture: Optimization formulation of network resource allocation; convergence analysis of primal and dual algorithms; Delay differential equations and applications to the study of congestion control algorithm; Interpretation of network architecture and algorithms in terms of optimization solution; Game-theoretic interpretation of optimization formulation and solution

• Mathematical tools: Markov chains and discrete-time queueing theory

• What happens at a link? Statistical multiplexing and large deviations

• Scheduling algorithms for switches and wireless networks: Maxweight scheduling, complexity, and distributed randomized algorithms, statistical physics techniques

• Mathematical tools: Continuous-time queueing theory; applications to connection-level models of the Internet; loss networks, heavy-traffic analysis

System Part

• P2P Overlay Networks

• Cloud Computing and Data Center Networks

• Wireless Sensor Networks

• On-Device Intelligence