CPSC662 -- Distributed Computing Systems (Spring 2002)

Slides  | Tests  | Homeworks | Resources | Projects (TA page)

In this course, we will study some of the most important concepts in the design of distributed operating systems. Throughout the course, we will try to relate the material covered to the support of large-scale computing on the Internet. We will tentatively discuss the following issues (a tentative guess for the allocated number of lessons per topic is given in brackets) :

Introduction [2]  Characterization of Distributed Systems  Network OSs vs. distributed OSs  Research and design issues  Interprocess Communication [2]  Issues in message passing  Client-server communication  Remote Procedure Call [1]  Design issues for RPCs  Case study: Sun RPC  Synchronization [2]  Event ordering / synchronization  Centralized vs. distributed schemes  Group Communication [4]  Ordered, reliable, and casual multicast  Group membership  Atomic group multicast  Virtual synchrony  Case study: ISIS  Service Replication / Reliable Services [1]

Distributed Shared Memory [4]  Shared memory Consistency models Design issues Case studies: Ivy/Munin/Treadmarks Distributed File Services [3]  Model  Case study: NFS  Case study: AFS  Case study: CODA  Modern Systems: OceanStore  Distributed Naming Services [2]  Names, addresses, routes, capabilities  Naming facilities, name distribution, name resolution  Migration  Supporting Internet Applications of the Future Issues Active networks, Liquid software, ... Peer-to-Peer, ... Other Cool Stuff...

As you see, most of the material served in the lectures is pretty basic. The idea is to provide a solid basis. We will complement this with a very strong hands-on component.

• " Distributed Systems: Concepts and Design; 3rd Ed"
  by G. Coulouris, J. Dollimore, T. Kindberg, Addison Wesley (ISBN 0-201-61918-0)

Reference Books

Below is a list of reference books. These books are thought to complement the textbook. Perusing them may help you better understand some issues discussed in class. More reference material will be added as we go along.

• "Distributed Operating Systems"
  by A. Tanenbaum, Prentice Hall
• "Distributed Operating Systems & Algorithms"
  by R. Chow and T. Johnson, Addison Wesley
• "Distribute Operating Systems; Concepts and Practice"
  by Doreen L. Galli, Prentice Hall

Assessment Method

The course will consist of a considerable number of challenging programming projects, two exams (first exam tentatively on March 5, second exam on April 25). Perhaps we will have a small number of homeworks or presentations.

The weight will be tentatively distributed as no less than 55% for projects and at most 35% for tests. (If we will have homeworks, the weights will be adjusted accordingly to account for the homeworks.)

Ideally, grades will be assigned along the following lines:

• A: Understanding of the material presented in class and in the literature. Excellent completion of machine projects. Indication of ability to independently develop new ideas based on material learned in class.
• B: Good understanding of material presented in class and in required reading. Good completion of machine projects.
• C: Some understanding of material presented in class and required reading. Completion of machine projects at unsatisfactory level.
• D: ....

Prerequisites

The official prerequisites are CPSC-410 or CPSC-463. I recommend you have a good idea of both.
I reserve the right to enforce the prerequisites, maybe even in form of a quiz at the beginning of the semester !

Projects

In this course we will not limit ourselves to talking about distributed computing systems, but will get our hands quite dirty. (Not as much as I would like, unfortunately.)

The projects have shown to be very demanding. If you do not know how to program at a sophisticated level, this is a good time to rethink your plans of taking this course! Experience with previous times this course was given has shown that you will not be able to learn the programming part as you go.

More information about projects will be posted soon.

The projects will be done in groups of two students each. This means that it is a good idea to start looking for a partner soon.

Things may change, so please check back later!

Late-Submission Policy

• All the submission deadlines will include extra time for the consideration of accidental events such as (not limited to) unavailable resource (machines are down,) sickness of students, stock market collapse, etc. This means that additional extensions are generally not granted.
• I reserve the right to penalize lateness with one third of the earned points of the item per class day.

Academic Integrity

It goes without saying that the highest level of academic integrity is expected for students in this class. While discussions among students on the analysis of problems and on the development of general solution approaches is welcome (encouraged, in fact,) the realization has to be of the individual student or of the individual project team. Spelled out, this means:

• For machine problems: Do not make design documents or code available to other teams. Do not use other teams' design documents or code. Do not make use of previous years' solutions.
• For homeworks: Do not make your solutions available to other students. Do not copy from other students' solutions.
• For exams: Do not use any information from other students. Do not pass any information to other students. By information I mean any information, not even the time-of-day.

Last But Not Least

Feel free to make comments, suggestions, and such things.