Network Performance

Course Description

An introduction to techniques for performance modeling and analysis of communication networks and computer systems. Primary emphasis will be on analysis of measurements, discrete event simulation and queuing theory. 

Prerequisites   
    Telcom 2300 (or other C/ Unix programming experience), Telcom 2100 (or other Introduction to networking course) and probability/statistics

Course Goals

• reinforced understanding of fundamental probability and statistics
• ability to effectively & correctly analyze data
• ability to design experiments & take meaningful measurements
• ability to design and program useful simulations
• an understanding of queuing theory and applications

R. Jain, The Art of Computer System Performance Analysis  John Wiley, 1991

Errors in the Jain book

Additional Reference
M. Malloy, Fundamentals of Performance Modeling, MacMillan, 1989
D. Claiborne, Mathematical Preliminaries for Computer Networking, John Wiley, 1994
F. Mazda, Analytical Techniques in Telecommunications, Oxford Press, 1996
A Pritsker, Introduction to Simulation and SLAM II, 3rd Ed., Halsted Press, 1986
A. Law & W. Kelton, Simulation Modeling and Analysis, 2nd Ed., McGraw Hill, 1991

Examples, homework assignments, notes etc. will be on the on the class schedule web page.

Web references
Virtual Laboratories in Probability and Statistics
The Statistics Homepage
Introductory Statistics: Concepts, Models, and Applications by David Stockburger
Probabilities in the Game of Monopoly®

• Measurement Related
• Network Monitoring Tools List of network management and measurement software
• CAIDA Measurement Tool Taxonomy
• Comprehensive list of IP measurement software/hardware
• Internet Traffic Archive
• IP Performance Metrics IETF Group
• Simulation Related
• Opnet
• Comnet
• CSIM
• JADE
• EZSIM(Free)
• NS-2simulator (Free)
• Example Animation using VINT simulator
• J-SIM (JAVA process-oriented simulator (Free)
• NIST ATM simulator (Free)
• C++SIM simulation package (Free)
• Web-Based Simulation
• Simjava
• Simtools Summary Discussion (somewhat biased) of simulation software tools
  
• SingleQ.c
  

• Queueing analysis
• Queueing theory resource page
• TK solver

• Homework ....................... 20%
• Midterm Exam ................. 25%
• Final Exam ....................... 35%
• Project ............................. 20%
  

• All work must be the student's own, unless collaboration is specifically and explicitly permitted. Any unauthorized collaboration or copying will at minimum result in no credit for the affected assignment and may be subject to further action under the Guidelines on Academic Integrity of SIS and the University of Pittsburgh. A document discussing these guidelines was included in your orientation materials, and you should acquaint yourself with them.
• Laboratory work will be done in groups, however individual lab reports must be submitted and be the work of each individual student.
• Assignments are due at the beginning of the class (see table). Late assignments may be folded, spindled, multilated or lost by the Instructor or GSAs before, during or after grading, unless specific arrangements are made by the student with the Instructor at least 3 days before the assignment is due. Surviving late assignments will be penalized.
• I reserve the right to reject any assignment that I deem unreadable. This means type up long blocks of text, use a spell checker and grammar checker. Write mathematical analysis neatly. Draw figures using a computer or at least a straight-edge. Everything must be on 8.5"x 11" paper, stapled together. You really don't want to know why this statement is in here do you?
• I reserve the right to modify course requirements administering surprise quizzes for up to 100 points if I feel that students are not pursuing a reasonable amount of course related reading, writing and computing. Should this option be exercised, the point distribution for grades will be changed accordingly.
• You are expected to check this page on a regular basis for any changes to the problems and syllabus. All are subject to change as the course proceeds and will be publicized in the updates section .
• The easiest means of contacting me is by electronic mail first ( jkabara@pitt.edu ). This helps me organize my time, and ensures a more accurate and efficient response outside of office hours. I read my electronic mail throughout the day every day and will usually respond with an answer within a few hours of reading it, and almost always the same day. If the nature of your question requires an interactive session, please come to my office during office hours as scheduled. If, for some unforeseen reason, I cannot make office hours, I will announce that in class whenever possible. I do not make appointments during office hours; I will be available on a first come, first served basis.
  
• You are expected to read the assigned material before the associated lecture
• Notes, assignments etc. are located as the links listed in the table, you are responsible for retrieving assignments from the web.
• Project
  
• A group (3 or less students)  course project involving a performance study using either measurements, simulation, queueing theory or a combination of these techniques  is required. Each group must send a  list of its members via email to the GSA and the instructor  by  no later than (see schedule). If  some individuals cannot find a group, I will appoint them to one. A suggested list of projects will be provided on (see schedule). Each group must select a project or propose a project (subject to approval) by (see schedule), if one is not selected I will select one for the group. An initial progress report is DUE (see schedule). The final project must be submitted no later than (see schedule).