Oracle8 Server Distributed Database Systems
distributed database, separate locations, synchronisation, two-phase commit, central database, replicated database, duplicated database, information systems
Knowledge Management: Emerging perspectives
Knowledge Management Network (including virtual library)
A distributed database is a single logical database, parts of which are located at more than one location or more than one server. It acts as a single collection of data but the database itself is geographically in separate locations. For instance a shop may store some of their data on site for their cash registers but other information within their database may be stored at head office but it can still be accessed by the store. This way all stores have access to up to date information and do not need to worry about wasting time downloading updates for their own computers. It can be updates centrally requiring only one source of data to be changed. Another part of this database may be located at the warehouse.
Distributed databases require synchronisation of locations so that all sites will receive any updated information. A two-phase commit is part of the DBMS used to maintain consistency across a distributed database. It checks whether both parts of the database are ready for transmission and then performs the transmission. If all of the transmission is not completed then none of the transaction will take place if it involves money.
A database may be centralised at the host computer, but accessible to remote locations. A "centralised database" accessable from remote locations used to be a fairly standard implementation. The downside was bandwidith limitations. This was why distributed databases were invented. A centralised databse is not a distributed database. One of the 12 principles of Distributed Databases as described by C.J. Date states that there is "no reliance on a central site" and that all sites must be treated as equals and that there should be no reliance on a central master site for some central service such that the system is dependent on that central site.
|
Part of the database at the central site may be duplicated at the remote site. The data chosen for duplication would, of course be related to activities at that remote location. In this scenario of "replicated databases" you need a whole raft of synchronsiation technologies and rules to maintain the database integrity. With this duplication of data, however, the advantage of reduced redundancy, which is one of the principal reasons for having a data base system in the first place, would be lost.
There may be complete databases at each remote location. Similarly "duplicated databases" can be useful in some instances (e.g when there is little change over time). This model is often used for publishing.
More On Distributed Databases: Distributed Databases, !2 principles of Distributed Databases
Related Topics (Information Systems) : DBMS, The 5 Types of Information Systems, Expert Systems, TPS, DSS, ES, MIS, OAS, Characteristics, computerised methods, non-computerised methods, flat file databases, relational databases, data modelling, URLs, storyboards, searching, HTML, backups, security methods, encryption, firewalls, virus' and virus scanners, passwords and security levels, social and ethical issues, data accuracy, data integrity, data validation, data bias, data reliability, privacy, ownership and control of data, access to data, data warehousing, data mining, freedom of information, acknowledgement of data sources, issues related to new trends, sequential access, direct access, distributed databases, storage media, querying, operators, wildcards, sorting, SQL, biometrics and ethical issues, face recognition
See Also: Knowledge Management, data mining,Information systems