A distributed database system refers to a logical collection of data that is shared between various users through a computer network. In distributed database systems, the storage devices are attached to different central processing units (CPUs). According to Ozsu and Valduriez, a distributed database system has collections of data that are separated into different partitions called fragments (2009).
Reasons why Organizations need Distributed Databases.Most organizations prefer to use distributed database systems because it facilitates decentralization of the organization through development of computer networks and spreading of processing units or sites.
Secondly, distributed database systems eases sharing of data hence easy accessibility of the data by all departments of the organization. According to Ceri and Pelagatti, distributed database systems also allow storage of data at specific places where it is frequently used (2010). This further facilitates accessibility of data. Distributed databases also encourage localized autonomy of various departments of the organization. This is because every department with a system site has full control of data within the site.
Thirdly, distributed databases improve the availability of data. This is because if data in one site fails, data on another site can be retrieved and used. Duplication of data within various sites or across the system also facilitates availability and reliability of data.
Fourthly, Riccardi (2011) argues that distributed database systems are economical to implement because it is cheaper to acquire smaller computers than single large computers for networking within the organization.
Last but not least, distributed databases allow for easy modification, addition, removal and manipulation of data stored within the systems without affecting the performance or activities of other users of the system.
Best Practices for Implementation of Distributed Database.For effective implementation of distributed database systems, the systems users, developers and administrators must interact adequately so that the various needs of the users are identified. The system designers and developers would then design appropriate systems that are capable of meeting such needs.
It is also important to note concurrent transactions within the system should not interfere with each other. The distributed databases should also have performance transparency that would enable it to have equal performance as the centralized database management system. According to Buretta, transparency means that users should have the capability of interacting with the systems as though they are operating a single system (2011). Moreover, data in the distributed database systems should have the ability to be accessed by users through local applications or global applications.
Software and Hardware Components of a Distributed Database System.A distributed database has database management systems (DBMS) software permits easy management of the distributed database and makes the distribution of data transparent to users. Data at each site of the database is under control of a database management system.
Additionally, the distributed system has localized applications that handle data automatically. It also contains a communication network that links sites with the distributed systems. A distributed database system also contains user request interface that accepts queries or requests for transactions from users. The distributed transactions manager translates requests from the user and converts them into actionable requests.
The synchronous software provides application integrity as well as reducing complexity of the distributed system. Furthermore, certain specialized software may also be used within the distributed database to facilitate replication and duplication of data.
On the other hand, the hardware components of a distributed database include computer storage devices, network servers such as intranet and extranets, data dictionaries and query languages.
Techniques for Distributed Data Placement.Various techniques that can be used for data placement in distributed databases include centralized data allocation, partitioned data allocation, complete replication and selective replication. Centralized data allocation is whereby a single database in one site is used to store data within the entire system. The data is then shared across computer networks. Partitioned data allocation refers to a database which has been divided into disjoint fragments or partitions. Each fragment is then assigned to one site. On the other hand, complete replication consists of maintaining complete copies of the database at each site. Lastly, selective replication refers to a combination of partitioning, replication and centralization techniques.
Riccardi asserts that data placement can also be enhanced through use of Local Area Networks (LAN) and Wide Area Networks (WAN) (2011).
Strategy for Distributed Query Processing.Query processing in distributed databases can be achieved through development of a centralized query processing unit. This would facilitate the ability of the systems to quickly respond to queries. This can be achieved through utilization of middleware such as SQL and XML.
Programmatic Access to a Database.An organization may allow for programmatic access to a distributed database through development of specialized tools such as MySQL and task-based representations to facilitate retrieval of data from the systems. For easy access through the internet, web-based database documentation tools that provide easy documentation of data may also be used. Similarly, a set of server-side programs that offer stable user interfaces for retrieval, manipulation and linking of data within the computer networks may be adopted.
The most appropriate distributed database system that would assist the national restaurant chain is presented in the diagram below. Sites 1, 2 and 3 represent various branches of the located in different places that are interconnected using the computer network.