COMP-614 Distributed Data Management

Winter Term 2010

News

Names and Numbers

Overview

Lecture Notes

Talk Schedule + Slides

Project Information

News:

Names and Numbers:

Lecture: Tuesdays/Thursdays 11:30-13:00, McConnell 103
Instructor: Bettina Kemme. Office McConnell 109N. e-mail: kemme at cs.mcgill.ca
Office Hours Instructor: TBA
TA: TBA

Overview

Distributed computing environments have become the standard IT environment. Examples are workstation clusters, mobile environments, the grid, publish-subscribe systems, multi-tier systems, or peer-2-peer database systems. Data management in such environments is challenging. In this course we will cover issues such as data replication, data caching, and distributed query execution. We will also look at fault-tolerance issues. We will first look at these topics from an abstract point of view, and then analyze them for specific computing environments, in particular for cluster-based and peer-2-peer environments. A common thread through most of these topics is that data must be kept consistent.

Possible Topic List

1) Principles in Distributed Data Management

Overview of communication mechanisms (point-to-point, RMI multicast, message queues, event-based, ...): 1-2 lectures
Transactions: Concurrency control (1-3 lectures), Atomic Commit Protocols (1-2 lectures)
Architecture Principles (1-2 lectures)
Replication Overview (1 lecture)
Caching Overview (1 lecture)
Distributed Query Execution: Principles (1 lecture)

2) Advanced Topics (for more information see Project Information

Cluster-based replication: consistency and query execution (2 lectures)
caching: consistency and dynamic caching (2 lectures)
Multi-tier architectures: performance, fault-tolerance, load distribution/load balancing (2-3 lectures)
peer-2-peer systems: search (file based and record based), replication (2-3 lectures)
distributed stream processing (1-2 lectures)
massively multiplayer games: zoning mechanisms, architecture alternatives (e.g. cluster-based, peer-2-peer) (2 lectures)
Security

This course is useful for all who want to

learn more about the essential problems associated with data distribution
learn the essential techniques and mechanisms to address these problems
deepen the knowledge in one or two sub-areas
get to know the current research questions raised in distributed data management
learn advanced working techniques that are essential to do research and/or get leadership positions in industry
- speak in front of an audience
- give small summaries (written or oral) about lectures and papers
- discuss with others
- read a paper or report analyzing its strengths and weaknesses
- do your own little research project

The course will be mainly given in form of lectures. I will introduce each of the topics with one lecture and then present advanced material. Each student has to give a lecture about a specific sub-area that includes material of recent research papers. Additionally, each student has to shortly present his/her research project at the end of the term. Topics will be posted soon.

Prerequisites: A course in database systems (e.g., COMP-421). Furthermore a computer networks (e.g., COMP-435/535) or distributed systems course.

Marking Scheme: The marking scheme will be as follows:

10% class participation: you are evaluated on how actively you participate in class discussions; pure attendance without saying anything will result in a B.
30% summaries and critiques: 1 oral summary, 4 written summaries, and 4 paper critiques:
- Summary: provides a 1-page summary of a past lecture/talk. It should contain the appropriate information such that a classmate who has missed that lecture/talk gets a good idea what the missed lecture/talk was about. For one or two of these summaries you also have to give a five minute presentation in class.
- Critique of a paper: contains a 1-paragraph summary of the paper, and then lists weak and strong points of the paper (weak points are more interesting). The papers to choose from are the papers other students present in their talks. It is expected that the paper critique is submitted BEFORE the corresponding talk takes place.
- Note that you can submit as many summaries/critiques as you want and the 5 best summaries and the 4 best critiques will make up your mark.
60% Term project. The term project is an in-depth study of one problem area. It consists of three deliverables:
- 15%: a survey report of 3-4 research papers,
- 15% own research work (implementation, evaluation or enhancements of algorithms)
- 30% a class presentation of some of the survey part (1 hour), and a presentation about the own research work (15-20 minutes)

Literature: The literature will be mainly based on survey and research papers. There are some books that might be interesting in the context of this course. Most of them are available in the library.

M.T. Özsu and P. Valduriez, Principles of Distributed Database Systems, 2nd edition, Prentice-Hall, 1999.
G. Coulouris, J. Dollimore, T. Kindberg, Distributed Systems 3rd edition, Addison-Wesley, 2000.
S. Mullender (ed.) Distributed Systems, 2nd edition, Addison-Wesley, 1993.
S. Abiteboul, P. Buneman, D. Suciu Data on the Web, Morgan Kaufmann, 2000.
K. Dittrich, A. Geppert Component Database Systems, Morgan Kaufmann, 2001.
P.A. Bernstein, E. Newcomer Principles of Transaction Processing, Morgan Kaufmann, 1997.
M. Buretta Data Replication, Wiley Computer Publishing, 1997.
G. Weikum, G. Vossen Transactional Information Systems, Morgan Kaufmann, 2002.
More will be announced

A note on academic integrity

McGill University values academic integrity. Therefore all students must understand the meaning and consequences of cheating, plagiarism and other academic offences under the Code of Student Conduct and Disciplinary Procedures (see http://www.mcgill.ca/integrity/ for more information).

French/English

In accord with McGill University's Charter of Students' Rights, students in this course have the right to submit in English or in French any written work that is to be graded.