Distributed Databases

A single Database can be located under one machine, which is called local database, or can be distributed under multiple machines controlled by one DBMS called distributed database.

Distributed Databases will be the choice due to many business and technological factors. It will be cheaper for large firms to have databases distributed over its offices, which might be in different countries or continents than having a central database, where through expensive dedicated networks connected to the central. Other reason might be the merger and acquisition trend that highly increased in the recent years, in such cases information resources should be combined into one, and this can be very challenging with huge centralized databases, running on different hardware and software platforms. Distributed databases can ease this process provided that the proper hardware and infrastructure is available. Another reason might be the need to serve customers in different geographical location, which in distributed database scenario a better response time can be achieved; like in ecommerce websites, which will increase performance. Some companies who might benefit from the distributed database are financial sectors, and airline reservations.

However, due to the high maintenance cost for the Distributed Databases, as high-level of expertise is required, in addition to the complexity of the design it might not be the always the best approach. For that I wouldn’t recommend to have a Distributed Database design if all computers are in the same location, this will cause a more complex environments with less benefits, as data reliability and availability can be maintained by database clustering which will be cheaper for such environments.

However, considering big organizations that have multiple offices located within the same country, distributed databases will be more beneficial and more feasible, as different departments and offices can have access to their database only and have more control on the data they relate to.  I would recommend the same if computers are spread worldwide, as communication will be more expensive, data integrity and reliability would be jeopardized if same data is kept locally in different locations.

Database mirroring would be used in case of disaster recovery sites for business continuity. Databases transactions can be duplicated to another copy of the database located somewhere else (different site, country,..) where the second database can be used incase of failure and only then it can be accessed directly.

Mohannad

References:

Bhargava, B.(1995) ’Building Distributed Database Systems [Online], Available from: http://www.cs.purdue.edu/homes/bb/cs542-06Spr-bb/building.pdf (Accessed 12 February 2011)

Bock, D, & Bordoloi, B. (2006), 'Trends in Data Management: 'High Availability.’’ Journal of Digital Information Management, 4, 2, pp. 150-154, EBSCOhost, [Online], Available from: http://ehis.ebscohost.com.ezproxy.liv.ac.uk/eds/detail?hid=109&sid=9fdabda6-c374-499f-a3ae-0bf919bbfdb7%40sessionmgr114&vid=3&bdata=JnNpdGU9ZWRzLWxpdmUmc2NvcGU9c2l0ZQ%3d%3d#db=iih&AN=21388775

(Accessed 12 February 2011).

Stollings, J. (n.d.) ‘Distributed Database Management Systems in the Modern Enterprise’ [Online], Avialable from: http://faculty.ed.umuc.edu/~meinkej/inss690/stollings/Stollings%20690%20Final.htm

(Accessed 12 February 2011)