1. Object-Relational and Object-Oriented Data Base MANAGEMENT SYSTEMS
Đoàn Văn Ban
Object-Relational and Object-Oriented Data Base MANAGEMENT SYSTEMS

2. Data Base Management Systems Traditional Database Systems
Chapter Overview
Data Base Management Systems
Traditional Database Systems
Object-Relational Database Systems
Object-Oriented Database Systems
The Future of Object-Oriented Database Management Systems

3. Data Base Management Systems: an Overview
Query
Object-Relational
DBMS
Relational DBMS
Object-Oriented
DBMS
No Query
The database guru Michael Stonebraker developed this chart to help describe the basic differences between the relational, object-relational, and object-oriented database management systems.
As he points out, one size does not fit all--each database system is best suited to different kinds of data and different kinds of applications.
First, in the lower left hand corner we start with simple data, that is, an arbitrary length sequence of characters, that requires no queries. A standard text processing system fits this model: you open the file, edit the file, and close the file. There is no need for SQL: Standard Query Language.
In the upper left hand corner we have simple data that a user wishes to query using the standard query language, preferably SQL-92, which recognizes datatypes such as text, integer, date, time, and memo.
Stonebraker is a developer of an object-relational database, and you can see his bias in mapping the object-oriented database as having “no query” capacity (an exaggeration--though historically a limitation with OODBMS).
File System
Simple Data Complex Data

4. Database management system
Definitions
Database
A collection of data records organized to facilitate the storage and retrieval of the data by information systems
Database management system
The software that runs a database
Almost all traditional database management systems are relational
Data are stored in tables
The primary key labels primary records
The primary key can be used to label records in other tables—foreign key

5. Traditional Database Systems (contd)
Information systems use a database management system for
Writing of data to disk and
Retrieval of data from the disk
Examples:
Mainframe: DB2, Oracle, Informix, Sybase
Personal computers: Microsoft Access

6. Traditional Database Systems (contd)
Execution of the database instruction

7. Traditional Database Systems (contd)
Database instructions can be embedded within many traditional programming languages
Including C and COBOL
Typical database instructions include
Retrieve
Insert
Delete, and
Update

8. Traditional Database Systems (contd)
The database management system manages the database itself
The programmer writes code to
Invoke the database management system, and
Provide it with data it needs

9. Object-Oriented Database Systems
A company with a traditional database management system is unlikely to use object-oriented database management systems for some time
They are still too complex
An extensive background in object-oriented database theory is needed
When an existing database is converted into an object-oriented database
The existing traditional information systems cannot access the new database
They have to be totally rewritten, at enormous cost

10. Object-Oriented Database Systems (contd)
Solution 1: Two databases
The one traditional, and
The other object-oriented
However, a major reason for using a database is to have just one copy of each record
Two databases means having
Two copies of all data, and
Two versions of every new information system

11. Object-Oriented Database Systems (contd)
Solution 2: A hybrid system
An object-oriented information system interfacing with a traditional database management system
Instructions are embedded in C++ or Java code
The database management system must
Manage the database (as before), and
Perform conversions from traditional to object-oriented systems, and vice versa

12. Object-Oriented Database Systems
When writing data to the traditional database it must
Extract the data from an object
Convert the data into a record, and
Write the record to the traditional database
After retrieving a record from the traditional database, it must
Insert the record into an object

13. Object-Oriented Database Systems
Solution 3: Object-Oriented Database Systems
The object-relational data base would contain the same data stored in the relational database.
It would also accommodate the following more complex data:
student photos
student voice print
Object-oriented model!
Object of class Student could extend an object of class Person
Subclasses Doctoral_Student, Masters_Student and Undergraduate_Student could inherit the methods of class Student
Objects of class Course

14. Object-Relational DBMS
Introduced in 1992 with the release of the UniSQL/X unified relational and object-oriented database system.
Then, Hewlett Packard released OpenODB (later Odapter), which extended its AllBase relational DBMS.
In 1993, Montage Systems (later Illustra): shipped the first commercial version of the object-relational Postgres.
SQL3 exists in draft form--anticipated ratification in July 1998.

15. Object-Relational DBMS Key Players
Informix: Illustra (Informix-Universal Server)
UniSQL
Oracle: Oracle8
Hewlett-Packard
Unisys: OSMOS
IBM: DB2 version 3
Sybase: Adaptive Server (introduced Sept. ‘97)
Just released--a Java-based ORDBMS:
Cloudscape’s JBMS
Note that Sybase just released an object-relational database, after years of development. Microsoft apparently has an object-relational database in the works.
Just this month, an Oakland-based start-up company, Cloudscape, relased the first “pure” Java-based object-relational database. Cloudscape’s developers are former employees of Illustra, Sybase, and Oracle.

16. Object-Oriented DBMS Suited for Complex Data
OODBMS allow better control of complex data and complex interrelationships among objects.
Examples:
Financial portfolio risk-analysis systems
Telecommunications service applications
WWW document structures
Design and manufacturing systems
Hospital patient record systems
The OODBMS integrates database capabilities directly into the object-oriented programming language (C++, Smalltalk, Java). This is different than SQL, which is a separate language that defines, retrieves, and manipulates data.
Client is the C++ program
Server is the OODBMS

17. Object-Oriented DBMS Developing Standards
The Object Database Management Group (ODMG): standards-setting body, organized in 1991
Rather than defining a completely new standard from the ground up, the ODMG standard builds upon the existing OMG, SQL-92, and ANSI programming language standards to define an application-portability framework among OODBMS
ODMG-93: the answer to SQL
ODMG-93: an extension of the CORBA Persistent Object Service
Three major components of ODMG-93:
Object Definition Language (ODL)
Object Query Language (OQL)
C++ and Smalltalk language bindings

18. Object-Oriented DBMS Standard: ODMG 2.0 (July 1997)
The Object Database Management Group (ODMG) has published the ODMG 2.0 Specification, a key milestone in the development of the only industry standard for object databases.
The ODMG 2.0 standard includes:
a new ODMG binding for Java that standardizes and simplifies the storage of Java objects in databases
a metamodel that provides a language-independent description of a database schema to tools and applications
improved C++ and Smalltalk language bindings

19. Object-Oriented DBMS Real-world applications
ObjectStore by Object Design
The Human Genome Project at the MIT Whitehead Institute (Cambridge, MA) is helping to sequence the entire human genome by the year They have two applications: MAPBASE (genetic mapping of mice) and LABBASE (physical mapping of humans and mice), both of which use ObjectStore as the data and model repository. They chose an OODBMS for the repository because of its ability to efficiently and quickly manage complex genetic data.
Monterey Bay Aquarium Research Institute uses ObjectStore to collect and store multi-media oceanographic research data (videos, image, etc.). They share their ObjectStore-based research data with other research organizations through an extranet application written in Java. They replaced Oracle with ObjectStore because ObjectStore reduced the amount of Java code required (no need to convert their objects to rows & columns) and because ObjectStore could be extended very easily to store any type of multi-media or textual data.

20. Object-Oriented DBMS Real-world applications
Versant Object Technology
NUSTAR International (Atlanta) uses Versant to accept, store, manage, and publish data about real estate in many markets across the country. OODBMS allows great flexibility for updating local editions--each of which includes photos and other information about homes for sale in a given area. Publications are updated very frequently. Company scans between 16,000 and 25,000 photos a week. Each photo is associated with a real estate listing that includes text and graphics (including maps, floor plans, or logos).
Some DBMS developers define limitations of the OODBMS as follows:
OODBMS are limited to small applications (small databases)
OODBMS don’t support a lot of concurrent users
It takes too much time to deliver an OODBMS application

21. Object-Oriented DBMS Limitations? A Response
1997 study of 24 applications found that:
15 applications occupied between 1 and 99 GB
Four applications supported more than 500 users, with the most common situation involving between 10 and 99 users.
Development can be time-consuming, but not in every case and delivery time is improving (facilitated by greater availability of vendor tools). Twelve to twenty-three months is average development period, but many companies delivered in less than a year.
Relational DBMS market will grow to $6.8 billion by 1997, and $8.1 billion by 2000 (compared to $2.9 billion in 1995)
ORDBMS market will grow to $1 billion by 2000 (compared to $14 million in 1995)
OODBMS market will grow to $430 million by 1997 and $600 million by 2000 (compared to $100 million in 1995)
Conclusion: Market for OODBMS is growing very fast, but still dwarfed by relational DBMS
Study conducted by Douglas K. Barry, the executive director of the Object Database Management Group (ODMG).
Note that no OODBMS studied comes close to supporting 30,000 concurrent users, as the NASDQ ORDBMS does.

22. Object-Relational DBMS or Object-Oriented DBMS? The Future.
Future databases (in perhaps 5 to 10 years) will likely be capable of everything.
Relational DBs will likely be replaced with object-relational DBs.
Object-relational DBs are quickly adapting the capabilities to deal with complex data (objects).
Object-oriented DBs are adding querying capabilities similar to those of relational DBs.
It is currently useful to distinguish data bases by their ability to deal with complex vs. simple data types and their ability to run a query.
There is some merit in each of the DB approaches. The ideal DB of the future will be able to do it all.

23. Future of O-O Database Management Systems
Object-oriented database management systems will become as easy to use as traditional database management systems
It will still be too expensive to convert existing databases and information systems to the object-oriented paradigm
However, new information systems will be written using the object-oriented paradigm, but will interface with the traditional database

24. Future of O-O Database Management Systems
New companies will use the object-oriented paradigm for both
Its information systems, and
Its database management system
Legacy systems are the biggest obstacle to the widespread utilization of object-oriented database management systems
Just as they are to object-oriented information systems in general
The object-oriented paradigm is the current choice for the development of new information systems
This will lead to the adoption of object-oriented database systems in the future

25. ObjectStore (www.odi.com), GemStore (www.gemstore.com),
OODMS Web Sites
ObjectStore (
GemStore (
Objectivity (
O2 (
Jasmine (
Versant (
POET (