SYLLABUS

CS 6411 OBJECT ORIENTED DATABASE MODELS AND SYSTEMS

(Fall 2000 - Navathe)

 

Objective: The objective of this course is to give a thorough understanding of the advances in data modeling, database design, and a new generation of applications that are a challenge for database management. We will emphasize the object-oriented modeling approach to support such applications. Mostly basic concepts, research papers, prototypes and approaches will be discussed. Students will have some exposure to commercial implementations. The course is readings oriented and a basic knowledge of database management and particularly of the traditional data models is assumed at the level of CS 4400 /CS 6400. Please see the instructor if you have not had any formal coursework in the database area.

Instructor: Prof. Sham Navathe
e-mail: sham@cc
Phone: 894-0537
Office hours: by appointment
Office: CoC 139
Secretary: Gwen Baker 894-8358, Room 263 CoC, gwen@cc
 Teaching Assistant: Wai Gen Yee. waigen@cc, Database Lab, Room 104A.
Class time: Tues, Thurs. 4:30-6:00
Class: CoC Room 102

Format of the course:

The course will be conducted as a series of presentations and discussions in class. The instructor will give the presentations for the first 4-5 weeks. After that the course will contain a series of presentations by student groups (size of group to be determined). The course traditionally focussed on what I have now defined as Part1: Basics of Object Modeling and Design. Under the semester system and with ongoing developments, I have expanded the scope. Students will do an assignment to show familiarity with some Object based DBMS and will do a group project or an individual research paper to explore a topic related to the Part II of the course. Take home quizzes and one or two take home exams will be given. There is no textbook and there will be no final exam.

Reference Books (almost all on reserve in the library) :

The following books represent the general body of knowledge related to object-oriented database management.

1. S. Zdonik, and D. Maier, (Eds.) Readings in Object-oriented Database Systems, Morgan Kaufmann, 1990.

A very good collection of readings that is getting outdated because of the rapid development in the field.

2. Won Kim, Frederick H. Lochovsky (Ed.), Object-oriented Concepts, Databases and Applications, Addison Wesley, 1989.

A collection of chapters by prominent authors covering important issues.

3. M. Brodie, J. Mylopoulos, J. Schmidt, (Eds.) On Conceptual Modeling: perspectives from Artificial Intelligence, Programming Languages, and Databases, Springer Verlag, 1984.

A thought provoking set of chapters addressing important conceptual modeling problems dealing with information representation and processing.

4. K. Dittrich, U. Dayal, A.Buchmann, (Eds.) On Object-oriented Database Systems, Lecture Notes in Computer Science, Springer Verlag, 1991.

Proceedings of the first major conference on the topic organized by the authors.

5. E. Bertino and L. Martino, Object-Oriented Database Systems- Concepts and Architectures, Addison Wesley, 1993.

A short book containing all relevant topics , but the coverage is rather brief.

Good discussion of query processing, storage management, and indexing.

6. R. G. G. Cattell, Object Data Management: Object Oriented and Extended Relational Systems, revised edition, Addison Wesley, 1994.

A good practical and basic introduction to the field. Covers standards and mentions a large number of prototypes.

7. F. Bancilhon, C. Delobel, P. Kannelakis (Eds.), Building an OODB System, the Story of O2, Morgan Kaufmann, 1992.

A detailed series of articles describing all aspects of the O2 system which became a commercial product (now sold by Ardent software).

8. A. Dogac, T. Ozsu, A. Biliris, T. Sellis (eds.), Advances in Object Oriented Database Systems, Springer Verlag, 1994.

Proceedings of the NATO workshop on OODB systems. (I have a chapter on the use of a classification based object model for database integration).

9. Michael Stonebraker and Dorothy Moore , Object Relational Database Systems, Morgan Kaufmann, 1996.

A comprehensive discussion of the object-relational approach.

Other Reference Books:

The following books represent methodologies in object oriented analysis and design which is a topic bordering database management and software engineering. Some of the books discuss specific applications using the object based approach.

1. Derek Coleman et al., Object-oriented Development, the Fusion Method, Prentice Hall, 1994.

2. Grady Booch, Object-oriented Analysis and Design with Applications, 2nd ed., Benjamin Cummings, 1994.

3. Haim Kilov and James Ross, Information Modeling: an Object-oriented Approach, PTR Prentice Hall, 1994.

4. Paul Harmon and David Taylor, Objects in Action: Commercial Applications of Object-Oriented Technologies, Addison Wesley, 1993.

5. Rajiv Gupta and Ellis Horowitz, Object Oriented Databases with Applications to CASE, Networks, and VLSI CAD, Prentice Hall, 1991.

PART 1: BASICS OF OBJECT MODELING AND DESIGN

We plan to cover a number of the following topics in this part of the course. It may not be possible to cover all of them. It is also dependent on the student interest as to how much we will cover ( based on what you turn in ).

TOPICS:

1 Advanced data modeling – A Survey of Semantic data Models

Fundamental work in data modeling is covered through surveys (one instructor’s own and another by Hull and King) and a discussion of functional, classification based and nested relational data models

( see paper list #1 – to be given out).

2. Basic concepts of Object-oriented Data Models

The main reference here is the O-O manifesto paper and chapters 10-12 in the book Fundamentals of Database Systems by Elmasri and Navathe. Many papers on concepts such as identity, abstraction etc. also exist.

3. Object oriented Analysis and Design

Popular methodologies such as Rumbaugh's, Booch's , Jackson's or the HP "fusion" approach. The most recent advances dealing with UML may be covered.

4. O-O Models and Systems

A large body of published material exists on research prototypes and commercial OODBMSs. e.g. ORION (from MCC) , GEMSTONE (from Servio) , IRIS (later: Open ODB - from H.P.), ODE (from ATT), O2 (now from Ardent), Jasmine (from Fujitsu, then C.A.). Other systems like Objectstore (from Object Design, now Excelon) or Poet have less written up in research literature.

5. The Object- relational approach

This approach is rapidly gaining in popularity. Ch. 12 in Elmasri/Navathe and the Stonebraker and Moore book cover this approach. It combines the strength of the relational DBMSs with the advantages of handling complex data in the Object-oriented systems. Students are encouraged to pick papers on other object-relational work and particularly on applications illustrating the utility of this approach.

6. Models and Languages that incorporate general programming functionality

There is a classical body of work on Database Programming languages (DBPL's); examples – TAXIS from U. of Toronto, Galileo from U. of Pisa..

7. O-O Query languages and interfaces

This topic has seen more work recently, particularly for better user interfaces and information retrieval. Classical work includes the work on ORION, and works of Jose Blakely, Bertino, Elliott Moss, Stanley Zdonik etc.

8. Object Algebras, Deductive capabilities and formalisms

This is a formal topic and needs more work. See works of Beeri, Lenzerini, Thalheim etc.

PART II : ADVANCED TOPICS IN OBJECT MODELING, DESIGN AND MANAGEMENT

This area is undergoing a lot of development and what is proposed below is just a list of topics for a start. The definition and scope will be revised as we go forward with this course and particularly based on the students’ input..

  1. Aspects of Object Data Management
  1. Implementation: deals with storage. clustering, and indexing and retrieval of complex objects.
  2. Special types of objects : documents, multimedia objects, engineering design objects, e-commerce product catalogs (issues of representation, integration, versioning, and schema evolution are important).

B. Current Research Related to Object Databases:

    1. Management of continuous data streams with related problems of object recognition, quality of service etc.
    2. Document management with documents as complex objects. Related problems of text-based retrieval and search engines as special indexing systems
    3. XML : basics, use of XML as a "standard", XML for integration, technologies related to XML.
    4. Object architectures, standards, and interoperability : current work on CORBA, COM/DCOM, EJB, Java Technology
    5. Systems to model and query unstructured data.
  1. Modeling of Emerging Applications
    1. Modeling the web and its implication on information querying and retrieval (pull technology)
    2. Modeling User behavior, user profiles, etc. and their relationship to personalization of web applications (push technology)
    3. Models specific to genome applications, simulation, mobile databases, and virtual reality.

NOTE: The above list of topics is by no means exhaustive. Students are welcome to propose additions/modifications to the above list of topics as they research the literature.

COURSE ORGANIZATION:

1. I propose to run this as an advanced course with group discussions rather than as a series of lectures. I do intend to present classical work in the area for the first 4 to 5 weeks or so. This will include a discussion of semantic data models and basics of object modeling for which a reading list is enclosed. I consider the semantic modeling area as the basis on which the O-O approach to data management is further developed. ( I am out of the country in the week of Sept 11 during which Leo Mark will discuss temporal modeling etc.).

2. For papers in list#1 two or three original sets will be made available with Gwen. You are asked to get a copy of all papers made (for your personal use).

Make sure to sign the checkout sheet and try so that several students can make a copy at the same time so as to improve the turnaround time. Do not check a binder out for more than half a day. (I will see if electronic copies of some of the papers can be made available to class for downloading).

3. During the above time, for the first 3-4 weeks, I would like you to get familiar with the OODBMSs available as free downloadable test copies on the web. The systems include: Objectstore (from Excelon. see objectstore.com), Versant, O2 (from Ardent) , Gemstone, Poet, Objectivity, Jasmine ( from Computer Associates). One of the assignments is to get familiar with an OODBMS, define and load and access a database and write a sample application. More on this later in the course.

  1. Beyond the first two topics in Part I. , mostly the students will be doing presentations of selected papers.

    2. For Part 1 : Each student must give me a list of at least 3 and up to 5 papers on one of the topics I mentioned or a related topic . Do not include a paper on the list without seeing the paper first. These papers/chapters must be outside of the above referenced books. A student should propose papers that "hang together" and belong to one or maximum two topics of interest to the student.

    DUE TUESDAY, 19th September:

    (a) Turn in a list of papers (hard copy) from which you could present one (or two - depending on the class size) in the class as follows:

    For each paper, give a complete citation in the same style as I have done on the list #1. Under each paper title, state in 3-4 sentences what you consider to be the "highlights" of this paper - content/contribution/uniqueness etc and to which category it belongs. I will try to pick one (or two) papers from your list.

    (b) Make sure that once I pick the papers to be presented, you have access to get a "clean" original copy of the paper. You have to provide either an electronic copy or a hard copy for all students of the paper you discuss in class ( we will batch these in some fashion).

    (c) I will then organize a reading list for the student presentation part of the course. I will try as far as possible to have you present a paper on your list - but that is sometimes difficult.

  2. Student presentations on Part1 will start Sept 28 or so and will go thru October. Every student will get to make one presentation on part 1 and one on part 2 in class. (depending on the class size, I will decide on the size of the groups). Each presentation must be accompanied by a set of transparencies – max. about 10 - copies to be made ahead of class time. Gwen can help you in making copies provided you turn in your stuff well ahead of presentation date. Each class will consist of two presentations.
    3.
  3. For part II : we will follow a similar procedure. List of papers due October 12th and presentations will be made during November.

Student's Responsibility:

    1. Presentation (s) to class. - judged by quality of presentation, the content on the transparencies, fielding questions, showing an understanding of the topic. I will keep attendance in classes when students give presentations.

(2) Take home exam - planned for the first set that I will be discussing.

(3) In class or take home quizzes: these will relate to the papers presented by other students. I will consult the presenters in making up these quizzes. (Every time you present a paper, you should email me a set of 3 or 4 quiz questions before your presentation.)

    1. OODBMS familiarity assignment: demonstrates that you can set up, access and retrieve data from some OODBMS. Details to be defined. Preliminary one page report on your findings of available OODBMSs due September 19th.
    2. PROJECT: to do some original research/implementation/application related to the topics of this course. If students have access to specific DBMSs, please discuss with me. We do not intend to install particular OODBMSs for this. I expect non-CS majors to do projects that are relevant to their own fields of specialty. (PROJECT PROPOSALS due September 26th ). Research papers must be individual. Implementation projects may be in groups. REPORTS/DEMOS DUE IN THE FINALS WEEK.

The grade breakdown will be about 40% presentations, 25% quizzes and takehomes, 25% project and 10% for class participation (also includes attendance at student presentations).

CURRENT DEADLINES:

Sept 19th : (i) List of papers for Part 1 , (ii) A writeup on available OODBMSs on the web. (format of the latter to be defined.)

Sept 26th : Project Proposal

Oct. 12th : List of Papers for Part 2.