Computer System Performance Modeling and Evaluation
                                                         Spring Term, 2004

Lecturer:          Prof. E. G. Coffman, Jr.
                         Office hours:  Monday 10:00 am to 12:00 pm by appointment. 
                                           Most other days will be
 available as well, including weekends.

                         Office: 813 Schapiro (CEPSR) Bldg.
                         Phone: (212)  854-2152 (office) (212) 316-9038 (home)  

Time:        Tuesdays,  6:10-8:00pm

Room:        Engineering Terrace 386

Syllabus:    The syllabus can be found in

Notes:        Math foundations can be found in  which is  a version of
                    a set of notes compiled by Philippe Nain.  The course notes will be supplemented by
                    many applications worked out in class.   Applications will be accumulated in the


                    Scheduling items:  On Mar. 9th, we will have a midterm.  
                    By Feb 24th, you must have a two-page proposal for a research project.
                    You may combine with one other person.

Course structure:
              There will be a mid-term, a final, and a research project.  Students may be
                  asked to give talks on their results.  Homework assignments will be given out
                  on a regular basis and maintained in  .     Problems in the Applications file
                  are to be done right after the applications have been discussed in class.

                           Review of course progress and preview of next week 

 Week 1:   Probability refresher, with elementary combinatorics/counting

 Week 2:  More probability, generating samples from a given distribution,

                     Poisson process, common distributions, limit theorems, inequalities

  Week 3:  Analysis of interleaved memory systems; expected value arguments,
                      Poisson processes in two dimensions (polling model)
                      Homework assignment: 2,3,4,5 from problem set.

   Week 4: Markov chains, expected value arguments, Little's theorem, carried
                      load = offered load.   The M/G/1 queue (for notes on the calculation
                      of the generating function see MG1.pdf  )

  Week 5:  Birth and death processes, analysis of the single server processor
                       sharing queue.  Proposals for projects due.

   Week 6:  Birth and death processes continued.  Models and problems of
                        fault tolerant computing.  Models of disk head motion.

   Week 7:  Return to continuous polling systems with a brief discussion of
                        continuous state-space Markov chains.  Expected value argument
                        for finding mean busy period durations.  General approach to
                        finding the transform of the busy period distribution.
                        Do problems 10,11,12,13 from the problem set to hand in 3/23

   Week 8:   Midterm to be administered by Teddy.  You will be able to
                         take the midterm home to work on it, but it must be handed
                         in within 24 hours.  (Put in my mail box or under my office

    Week 9:   Semester break

    Week10:  Priority queues, foreground-background service disciplines,
                          brief review of midterm.  Please set up appointment to discuss
                          progress on the project.