Reference Architecture of DDBMS

Centralized Database System:

A type of database system in which all system components reside at a single computer or site is known as a Centralized database system. The system components include database and DBMS etc.

Distributed Database System:

A type of database system in which the database is stored physically across computers or sites at different locations is known as a distributed database system.
The sites of a distributed database may be spread over a large area connected via a wide area network (WAN) or over a small area such as a building or a campus connected via a local area network.

Decentralized Database:

A type of database system in which the database is stored on computers or sites at different locations is known as a decentralized database system. In a computer decentralized database, the computer system is not interconnected via a network. So the users at various sites cannot share data of any type. It is a collection of independent different types of databases.

Distributed DBMS:

The users access the distributed database via applications. In a distributed database system database, the application running at any of the system's sites should be able to operate on any of the database management system (DBMS).

Advantages:

Different advantages of the distributed database are as follows:

• Local Control
• Reliability and Availability
• Efficient Querying Processing
• Modular Growth
• Economics

Disadvantages:

Different disadvantages of the distributed database are as follows:

• Software Complexity & High Costs
• Processing Overheads
• Data integrity
• Complex Database Design
• Slow Response 

Functions of DDBMS:

A distributed DBMS provides access to data at various sites. A DDBMS is required to perform the following functions in addition to the function of a DBMS.

• Extended communication services
• Extended system catalog
• Distributed query processing 
• Extended concurrency control
• Extended recovery services

Types of DDBMS:

A DDBMS can be classified as follows:

Homogeneous

The term homogeneous database means that the database technology is the same or at least compatible at each of the locations. The data of any type at the various locations are also compatible. 

Characteristics of Homogenous DDBMS:

• Data redistributed across all the nodes.
• The same DBMS is used at each location in it.
• All data are managed by the distributed DBMS in homogenous so there is no exclusively local data of any type.
• All users access the database through one exclusive scheme or database definition.
• The global scheme is simply the union of all the local database schemas.

Heterogeneous Database:

In heterogeneous distributed potentially different DBMS's are used at each node. It is difficult in most organizations to force a homogeneous environment, yet heterogeneous environments are much more difficult to manage.

Characteristics of Heterogeneous DDBMS:

• Data are distributed across all the nodes.
• Different  DBMS is used at each location
• Some users require only local access to the database

DBMS Transparency & Gateway:

The term DBMS transparency refers to the ability to hide the knowledge that the local DBMS may be different. It applies to a heterogeneous environment. It is one of the most difficult transparencies to provide DBMS.

Object-Oriented Databases:

Object-Oriented Programming (OOP) methodology has altered the way we program. Object-oriented ideas are also influencing how we think about other concepts such as databases. The merging of object-oriented programming and database management systems is the direct result of the increasingly difficult task of dealing with complex data requirements and the issues of modeling real-world data with relational database technology.

• Entity Sets and Tables:

OODM's idea of class resembles the ER model's idea of an entity set or table. As an object, OODM classes are more powerful than the ER mode idea of an entity set or table. A class not only describes data structure but also describes the behavior of the class objects.

• Relationship:

An important property of a model is how they represent relationships among different components of data. OODB has two types: inter-class or class hierarchy. This is constructed against the ER value-based relationships.

• Versioning:

Another difference between traditional and E-R based databases are the OODBs support of versioning. Just as a document can have multiple versions so can a database. One can load a system and change aspects to determine how this would affect the overall system. Then the original system can be reloaded intact.

• Data Access:

Traditionally, data access can e viewed as a two-level storage model. It needs a link between the main memory and secondary storage device to access data contained in the database. The user needs data to analyze and use it. To get the data, The user makes a query through SQL. SQL accesses the data on a secondary storage medium. It then transforms and types checks the data before it transfers the data to the user.

• OODM + DBMS = OODBMS:

An OODBMS actually manages the data access and the querying of the data from the databases. It manages multiple users typing to access the data at the same time. It provides recovery services and manages all the databases in its system.

The "Object-Oriented Database System Manifesto" was introduced at the First international conference in Deductive and Object-Oriented Databases in Kyoto, Japan. It provided the following musts for an OODBMS:

• OODBMSs should support complex identity.
• OODBMSs should support object identity.
• OODBMSs should support types or classes
• OODBMSs classes nor types should inherit from their ancestors.
• OODBMSs should be extensible.
• OODBMSs should manage a very large database.
• OODBMSs should recover from hardware and software.
• OODBMSs should have a simple way of querying data.

Advantages of OODBMS:

• OODBMS's provide many advantages. These advantages are important because they solve many of the problems traditional systems cannot solve.
• First,, the amount of information that can be modeled by an OODBMS is increased, OODBMS's provide complex objects that allow ease of integration for multimedia, CAD, and other such specialized databases.

• In an OODBMS, versioning is available to help model various changes to systems. With versioning, one would be able to revert to previous data sets and compare the current sets to the previous.

Disadvantages of OODBMS:

In addition to modeling advantages, OODBMS has system disadvantages as well.

• Traditional ER systems have been used for a long time and a change would stray from established ideas. It would require people to think differently, and in some cases, relational users lack the OO foundations needed to work with OODBMS's
• Educating people on OO foundations is a very difficult process.
• Because of the change, even more, time would be required to move the data into the new OODBMS's.
• Traditional systems and OODBMS must understand each other and the relations they are representing.
• There is no query language such as SQL. While it is easier to make complex queries with OODBMS's, no query language exists.
• There are no standards for design and implementation in place.
• OODBMS can solve problems of traditional systems, but a standard is required.