Computer Science Department - College of Arts & Sciences

Fundamentals of Computing Systems

CS-350: Fall 2016

Instructor

    Name: Prof. Azer Bestavros
   Email: best@cs.bu.edu
  Office: MCS-175, 111 Cummington Mall (ext: 3-9726)
   Hours: Mon 09:30am-11:00am
          Tue 05:00pm-06:30pm          
          or by appointment arranged through email

Teaching Fellow

    Name: Qiaobin Fu
   Email: qiaobinf@bu.edu
  Office: Undergrad Lab Room 302 at 730 Commonwealth Ave
   Hours: Mon 01:00-3:00pm
          Fri 03:00-4:00pm
          or by appointment arranged through email

Lectures

   Place: SED-130
    Time: Tue/Thu 3:30pm-5:00pm

Sections

   Place: See below
    Time: Wed 09:00-10:00am (in MCS B25)
          Wed 10:00-11:00am (in MCS B21)
          Wed 01:00-02:00pm (in MCS B19)

Important Dates

          Last day to drop the class without a W grade is Wed 10/12
          Last day to drop the class with a W grade is Thu 11/10
          Midterm Exam dates are Thu 10/13 and Tue 11/22
          Final Exam date is Thu 12/20 @ 3pm

Cyberspace Coordinates

     URL: http://www.cs.bu.edu/~best/courses/cs350/F16
  Bboard: http://learn.bu.edu/
  Piazza: http://piazza.com/bu/fall2016/cs350
Facebook: http://www.facebook.com/groups/bucs350
 

Schedule of Lectures and Assignments

Date

Lecture Topic

Assignments

9/06

Background: System Abstractions and Performance Metrics

 

9/08

Background: Elementary Probability Analysis

#1 - Due 09/20

9/13

Background: Probability Distributions and Expectations

 

9/15

Performance Evaluation: M/M/1 Queuing Model and Analysis

#2 - Due 09/27

9/20

Performance Evaluation: Discrete Event Simulation

 

9/22

Performance Evaluation: Measurement & Confidence Intervals #3 - Due 10/04

9/27

Performance Evaluation: M/M/1 Variations & Queuing Nets

 

9/29

Performance Evaluation: Use Cases

#4 - Due 10/18

10/04

Resource Management: From M/M/1 to GPS  

10/06

Resource Management: Basic CPU Scheduling 

#5 - Due 10/25

10/13

Exam #1  

10/18

Resource Management: State-Sensitive (I/O) Scheduling

 

10/20

Resource Management: Real-Time & Priority Scheduling

#6 - Due 11/01

10/25

Concurrency: Mutual Exclusion and Dekker Algorithm

 

10/27

Concurrency: Peterson and Bakery Algorithms

#7 - Due 11/08

11/01

Concurrency: Semaphore Synchronization

 

11/03

Concurrency: Producers/Consumers & Readers/Writers

#8 - Due 11/15

11/08

Concurrency: Deadlock Management

 

11/10

Concurrency: Transactional Semantics #9 - Due 11/22

11/15

Concurrency: Global Snapshot and Lamport Clocks  

11/17

Concurrency: Networking and RPC Abstractions  

11/22

Exam #2  

11/29

Distributed Systems: MapReduce Programming Abstraction #10 - Due 12/08

12/01

Distributed Systems: Hadoop & Cloud Computing Platforms  

12/06

Distributed Systems: Advanced Topics  

12/08

Wrap-Up  

12/20

Final Exam (3:00pm-5:00pm)

 

Note: Changes to the schedule will be announced in class if need be

Important Dates:

  • Last day to drop the class without a W grade is Wednesday 10/12/2016

  • Last day to drop the class with a W grade is Thursday 11/10/2016

  • Exam dates on 10/13/2016, 11/22/2016, and 12/20/2016 are immutable


Overview

This course is a required sophomore/junior-level Computer Science course. It covers the fundamental concepts underlying the design and implementation of computing systems.

The philosophy underlying the design of this course is that students should be familiarized with problems that reoccur in software systems, and should be acquainted with the set of classical algorithms/techniques for solving such problems. In particular, it is important to develop the ability to recognize standard problems in different wordings and within unusual context, and match them with appropriate solutions.

Catalog Description:

Rigorous treatment of invariant concepts, algorithms, and performance evaluation methods underlying computing systems design. Topics include modeling and analysis of concurrent processing, computational resource scheduling and consumption, and performance evaluation techniques.

Prerequisite Courses:  
    CS-112 and CS-210: Programming and basic software/hardware interface concepts.
    CS-131 or CS-237: Elements of discrete and continuous probabilistic analysis.

Recommended Follow-up Courses:

The following courses build on knowledge assimilated in CS-350. They are natural follow-up courses in decreasing order of relevance (more or less).

    CS-451: Distributed Systems and Cloud Computing
    CS-552: Introduction to Operating Systems (and CS-553 as a follow-up to that)
    CS-455: Introduction to Networking (and CS-556 or CS-559 as follow-ups)
    CS-470: Performance Evaluation and Modeling of Systems and Networks
    CS-410: Software Systems and Systems programming
    CS-460: Introduction to Database  (and CS-562 or CS-565 as follow-ups)
    CS-450: Introduction to Computer Architecture (and CS-550 as follow-up)

Function in the Curriculum:

As a core undergraduate course, CS-350 does not focus on particular implementations or specific technologies. On the contrary, it stresses the fundamental concepts and basic algorithms that have survived (and are likely to survive) the evolution of computer software systems in general, and operating systems in particular.

This course is significantly different in purpose and coverage from CS-210 (Computer Systems) and CS-410 (Software Systems). CS-210 and CS-410 introduce computer systems to sophomores using a hands-on approach by examining the "mechanics" and/or the "plumbing" of modern (or typical) computer operation through exposure to various interfaces between architecture, compilers, loaders, linkers, and run-time systems. CS-350 does not focus on a particular interface or a particular technology, rather it deals with fundamental notions and algorithms that are common to computing systems in general and to software systems in particular.

This course is required or strongly recommended for the following advanced courses: CS-552 (Advanced Operating Systems), CS-550 (Advanced Computer Architecture), CS-551 (Parallel Computing Architectures and Models), CS-455/655 (Data Communications), CS-460/660 (Databases), and CS-470/670 (Performance Analysis). In that respect, this course covers basic notions that we expect students (graduates and undergraduates alike) enrolling into these classes to have mastered. For example, it is inconceivable that a student who is interested in taking a database or an operating systems class, not to have been exposed to the notion of concurrent processing and mutual exclusion. Similarly, it is unconceivable for a student enrolled in a networking, database, or operating systems class, not to have been exposed to performance analysis using simple queuing systems, for example.

Topical contents:

The topics covered in this class could be grouped under 3 general themes:

  • Performance Analysis and Evaluation (with probability refresher)

  • Scheduling and Resource Management

  • Concurrency and Synchronization

The first half of the semester will be devoted to the coverage of the first two of the above themes, whereas the second half of the semester will be devoted to the coverage of the third of the above themes.


Class Meetings and Communication Forums

Lectures:

Attending lectures is mandatory and will account for 10% of the final class grade. To take attendance, frequent one-question quizzes (at some point during a randomly selected set of lectures) will be given.

You are responsible for ALL the materials covered in the lectures including any topics not in the notes or references.

Class participation and questions are very welcome during the lectures. There is no such thing as a "stupid question". Failure to ask a simple question might result in the inability to follow the rest of the lecture (and even following lectures).

Up to 10 minutes at the beginning of every class may be devoted to answering questions related to topics discussed in previous lectures. This is a particularly good time to ask if you have questions.

Discussion Sections:

If you are taking this class then you have signed up for a weekly one-hour discussion session. Please check the course web page for the coordinates of your teaching fellow and of the place/time of the discussion sections.

The teaching fellow will be leading the discussion sessions. Materials covered in the discussion sessions will be elaborations on materials covered in the lectures. In some of the discussion sections (especially at the beginning of the course), the teaching fellow will be covering background/prerequisite materials that many students typically do not master. Also, the discussion sessions will be the venue used to answer any questions (or provide clarifications) regarding the homework assignments.

If (for some reason) you miss (or cannot make) the discussion section for which you are signed up, then please make sure to attend another one in the same week.

Office Hours:

You should come to the Instructor and/or Teaching Fellow's office hours with your questions well-thought out. Office hours are NOT meant to be tutoring sessions; they are meant to answer specific questions about the material covered in the lecture, discussion section, and lecture notes.

Questions like "could you repeat your explanation of ..." or "I do not understand section ... of the lecture notes" should be asked in the lecture and/or discussion sections.

On-line Forums and Facebook Group Usage:

The class communication forums (e.g., Piazza, Blackboard, and Facebook) are for you to use as channels for asking questions and seeking clarifications, whether from the instructor, teaching fellow, or classmates. However, like any broadcast medium, you should be careful not to abuse these channels. This means that "one-liners" and "cute" commentaries should be avoided.

You will be held responsible for any inappropriate postings on these forums.

Participatory Credit:

In an effort to make things a bit more dynamic during lectures and sections, the instructor and the teaching fellow may occasionally award "Bonus" credit points for student participation. Also, as a further motivation to read the notes thoroughly, bonus points will be awarded to students who discover technical "bugs" in the lecture notes, or who provide significant contributions. Such bonus points will count as extra credit points for the homework that is due next and will be capped at 10 points per homework. Any points above the maximum grade for the homework will not count (i.e. if your grade in a homework is 100/100 and you get 10 extra points, your grade for the homework will continue to be 100/100).


Class Notes

[W] Azer Bestavros, Lecture Notes on Fundamentals of Computing Systems. These notes are a book "in-progress", and hence they are subject to change. They constitute THE required readings for the class.

Note: These Lecture Notes are available from the Class Home Page. They are all in PDF format, so you will need access to an Acrobat Reader.


Homework Assignments

Homework assignments constitute an important part of this course. They are designed to help you understand the materials covered in lectures and in assigned readings. It is only by doing the homework that you really learn the material.

Nature, Frequency, and Grading:

There will be 10 homework assignments handed out throughout the semester (subject to change).

Homework Assignments will be worth 30% of the final grade. In calculating the overall homework grade, the homework with  the lowest grade will be dropped; i.e., the nine homework assignments with the highest grades will count toward the 30% of your class grade. So, each assignment is worth 3.33% of your class grade.

Homework assignments are not necessarily weighted equally and problems within homework assignments may also vary in their relative weights. Harder/longer homework assignments will constitute a larger percent of the combined grade for all homework assignments.

Some of the problems in the homework assignments will require programming. At least two of the homework assignments will require writing a discrete event simulator for a queuing system, which (for some) may prove a bit challenging given the "different" nature of such a program (management of queues, events, state of the queuing system and of the simulation, etc.)  Also, a number of homework assignments will require programming using threads and/or using a cloud platform such as Map/Reduce on Hadoop. While you are free to use any programming language or environment that you deem appropriate, the recommended language is Java and the recommended programming environment is Eclipse.

Homework assignments may be fairly long, each taking an average of  4 hours to complete (assuming that you are caught up with readings and are fluent in programming), although the time required will vary quite a bit from student to student and assignment to assignment. Always allow more time than you think you will need!

Electronic Submission Process:

Homework assignments will be posted on-line at the end of the week (typically on Fridays). As a tree-friendly practice, no hardcopies will be distributed or made available.

Completed homework assignments are to be handed in electronically, typically using gsubmit ( http://www.cs.bu.edu/teaching/hw/gsubmit/). Check details regarding submitting solutions and code. Typically, homework assignments will be due before lecture on Tuesdays (12 days after being handed out).

If your homework solutions (or parts thereof) are handwritten, you will need to scan your write-up and upload that electronically. Scanning capabilities are available at various university libraries and also through the main CS office on the first floor of MCS (during normal business hours, office staff can help you email a scanned copy of your work to yourself). Please take this into consideration so that you don't get stuck having to do this at the last minute.

Graded assignments will be returned (also on-line) in a timely fashion. It is the student's responsibility to make sure that the grade they received for each homework is properly recorded by the instructor and/or teaching fellow. This can be done by periodically checking your grades on-line through the course's intranet. 

Late Policy:

For full credit, your homework must be submitted electronically by the deadline. There will be a hefty penalty of 25% for a homework submitted one class late. You may submit part of the homework by the deadline to avoid the penalty on that part. No homework will be accepted if late by more than one class. There will be no exceptions to this policy, except for certified medical excuses. In such cases, extensions will be granted only if (and until) the homework solutions are posted (hopefully, about 1 week after the original due date).   


Exams and Quizzes

Attendance Quizzes:

Throughout the semester, in random lectures, I will be giving one-question quizzes. Each of these quizzes will be given at some point during the lecture (not necessarily the beginning or end). Each will consist of a single (easy) question on the material covered in that lecture. The main purpose of taking attendance and of these quizzes is to be a measure of attendance, since I expect anybody who attended the class (and who was not asleep) to be able to answer the question. You are allowed to miss one attendance quiz without being penalized.

Attendance quizzes will go towards the 10% of the final grade for class participation. 

Interim Exams:

There will be two in-class exams. These exams (combined) will be worth 30% of the final grade, and will cover the material presented from the beginning of the semester (or from the previous in-class exam) and up to the last lecture before the exam. Please check the class schedule for the specific dates of these exams.

Final Exam:

The course final exam is worth 30% of the final grade, and will cover the material offered throughout the semester. Please check the class schedule for date and time. The place for the final exam is typically the same as that of the lecture (and will be announced in due time, if different).

Policy on Missed Exams:

Please mark the exam dates on your calendar (and remember them when you make your recess and end-of-semester travel plans!) There will be absolutely no make-up exams, except for medical emergencies. For medical emergencies, you must provide a letter from a doctor, specifying the period of time during which you were unable to attend an exam.


Grading

Grade Breakdown:

The final class grade will be broken down as follows:

  • 10% on Class/Piazza Participation

  • 30% on Homework Assignments

  • 30% on Interim Exams #1 & #2

  • 30% on Final Exam

Grading Process:

Grading (except for the final exam) is done by a number of class graders, under the direct supervision of the Teaching Fellow(s). If you have an issue with a grade (homework or exam), please contact the Teaching Fellow(s). If your issue is not resolved, then (and only then) please contact me. In doing so, please note that (to ensure fairness and grading consistency) it is seldom the case that the I will overrule a Teaching Fellow.

This class is not graded on a curve, i.e., there is no prescribed proportions for specific grades. This means that if everybody's performance in the class deserves an A, then everybody will get an A. The converse is also true! Therefore, don't be satisfied with an "average" grade because that average could well be less than what you expect.  Being a "gateway" to all other (more advanced) CS "systems" classes, you should expect this class to be competitive. Thus, make sure you work hard from the very beginning.

In previous years, the average grade for this class was around B/B-. The minimum grade for this course to count towards the CS concentration is C.

An instructor is not allowed to give W (withdrawal) grades. One can get such a grade only by dropping this class by the deadline specified by the registrar office for withdrawals with or without a W grade (check the registrar's office calendar for the exact date). Also, an instructor is not allowed to give an I (incomplete) grade except if a student misses completing assignments and/or misses taking tests due to circumstances beyond their control.


Academic Code of Conduct

In addition to the normal and well-understood strictures against cheating on exams, altering transcripts, and etc., there are other varieties of academic misconduct described in the BU Academic Conduct Code and which you must be aware of when working on assignments.

The most relevant section of the Academic Code of Conduct for the assignments in this class relates to Plagiarism.

Plagiarism. Any attempt by a student to represent the work of another as his or her own. This includes:

  • Copying the answer of another student on an examination or copying or substantially restating the work of another person or persons in any oral or written work without citing the appropriate source, and collaborating with someone else in an academic endeavor without acknowledging his or her contribution.

  • Knowingly allowing another student to represent your work as his or her own. 

  • Submitting the same work in more than one course without the consent of the instructors.

Policy on Collaboration:

There is nothing wrong in principle about discussing the topics covered in the course with your friends and colleagues. However, there are severe consequences to plagiarism. In particular, when you submit a homework with your name on it, you are claiming that the work contained therein is your own; it is Plagiarism to submit work under your own name in which:

  • You collaborated with another student (current, former, friend, etc.) on homework problems;

  • You copied the solution from another student; or

  • You obtained the solution from a book or other sources.

The above is nothing less than plagiarism and will be punished accordingly.

If you are in a situation whereby you have collaborated with a student on a homework problem, or obtained homework solutions (or parts thereof) from some other source, then you should clearly and unambiguously disclose this in your write-up. Failure to do so constitutes plagiarism.

If you are caught cheating in a test or plagiarizing a homework, you will automatically receive an F and the matter will be reported to CAS Academic Conduct Committee, who recommends sanctions to the Dean of the College.

Handing in your own assignment a day or two late will affect your grade far less than turning in a copy of someone else's work on time!


Teaching Staff

Azer Bestavros (PhD 92, Harvard U) a Professor in the Computer Science Department at Boston University, which he joined in 1991 and chaired from 2000 to 2007. He is the Founding Director of the BU Hariri Institute for Computing, set up in 2011 to support "a community of scholars who believe in the transformative potential of computational perspectives in research and education." His notable research contributions include pioneering studies of web push caching through content distribution networks, self-similar Internet traffic characterization, game-theoretic cloud resource management, and safety certification of networked systems and software. Funded by over $25M of grants from government and industry, his research work yielded 18 PhD theses, 7 issued patents, 2 startup companies, and hundreds of refereed papers that are cited over 16,600 times. In 2010, he was named BU United Methodist Scholar Teacher in recognition of "outstanding dedication and contributions to the learning arts and to the institution."

 

Qiaobin Fu is a third-year PhD student in the Department of Computer Science at Boston University. He is a member of the Networks Research Group, and his advisor is Professor John W. Byers. Currently, his research interests center on designing algorithms and building systems in Networking. He received a B.Eng. from Dalian University of Technology in 2011, and an M.S. from University of Chinese Academy of Sciences in 2014.