Distributed Database Management Systems

Introduction

Mostly we come across a concept of DBMS, but a very few of us have a knowledge of what it actually means. Let us have an overview of this topic in a very simple and understandable language. A distributed database is a collection of logically inter-related databases that are spread physically across multiple computer network connected by a data communication link.

Distributed Database Management System is a software system that permits the management of distributed database and provides an access mechanism that makes the distribution transparent to the users, who perceive the database as a single database.To understand this concept in a better way, lets have a look on the comparison between Distributed DBMS against the Parallel DBMS:

Distributed versus Parallel DBMS

Distributed DBMS	Parallel DBMS
A distributed database contain data in two or more server computers at separate locations	A typical parallel database contain data as one set of files in one location, because several computers share the workload
A distributed database server executes all the steps itself.	A parallel database assigns and allocate steps to different computers
DDBMS can be connected by using network like internet	It can be connected by using super speed LANS and even with switches
In Distributed databases each office has its own database with improves the access.	Parallel database does not improve access to remote locations
If a problem develops and one fails, outage only effects the users using the failed computer as the remaining local servers continue to process information	If a problem develops and one fails, the server program removes it from service but the database remains available as parallel database. However the performance is reduced monitoring the state of each computer
Machines are far from each other like in different continent	Machines are physically close to each other like same server room
Usually Communication cost and problems are high as it can never be ignored	Communication cost is low
Distributed databases usually use shared-nothing architecture	Parallel database use shared-memory, shared-disk, or shared-nothing architecture

Advantages and Disadvantages of Distributed DBMSs

Following are the advantages of DDBMs:

·         Distributed database management system provides aid to locate database over several locations at a time.

·         Database can be retrieved over different networks anytime.

·         Distributed database management system is secure.

·         Data can be linked and updated from different tables which are placed on different machines.

·         It is much reliable as if a certain node fails, the integrity of the distributed database is maintained and system allows it to reroute the request of failed node to another system.

·         Distributed database management system has a share ability because users have control of data as it is placed locally.

·         Distributed database management system is flexible as we can add new data without affecting the operations of old data.

·         Distributed database management system is much faster as it allows parallelism which speed up the data to perform certain operations at a time.

·         Distributed database management system is more cost efficient.

Disadvantages of distributed database:

Following are the disadvantages of DDBMS:

·         Optimization of database is challenging in a distributed database management system because of the complexity of the system.

·         There is no specified data format for handling distributed database management syste.

·         Designing of distributed database management system is more difficult due to fragmentation and replication of data.

·         Increased complexity requires much maintenance costs for Distributed Database management system.

·         Performance is reduced as the data is accessible from a remote system.

·         Handling of deadlocks in Distributed Database management system is much difficult.

Classification of DDBMS

A DDBMS may be classified as homogeneous or heterogeneous.

Homogeneous DDBMS:

In a homogeneous distributed database:

·         All sites have identical software to deal all types of requests from users.

·         Are aware of each database in the system and work itself by processing user requests.

·         Each site yields autonomy in terms of right to alternate schemes or software.

·         Appears to user as a single system which allows increased performance by using parallel processing.

Heterogeneous DDBMS: 

In a heterogeneous distributed database: 

- All sites have different software to deal all types of requests from users

- Sites may not be aware of each database in the system and provide only limited facilities for processing user requests  

- Each site may use different schemes or software

- If both the hardware and software are different, then translation is required at both sides which results extremely complex processing

Summary

To link the individual databases of a distributed database system, a network is necessary. It facilitates inter-database communication across a network. It connects users and database system to communicate with each other and support remote and distributed transactions in a distributed database. Whereas networking protocols unite all the processes and controls of initiating and
accomplishing communication between different database network systems enabled devices. So basically DDBMS are expected to provide backup and also data recovery by database administration, maintain consistency of replicated data by extended concurrency control, include query optimization as by Distributed Query Processing, including providing access to remote data, providing extended recovery of data at failure of communication links, providing access and privileges to the distributed data by extended security control, extending communication services to provide access to remote sites and allow the transfer of queries and data among the sites using a network. Also providing transaction management by extended system catalogue to store data distribution details.
You can share your feedback below if you want us to keep you updated with such short yet informative academic topics.