A Survey on Distributed Databases Fragmentation, Allocation and Replication Algorithms
H. A. Fuaad *
Thamar University, Yemen.
A. A. Ibrahim
Thamar University, Yemen.
A. Majed
Thamar University, Yemen.
A. Asem
Thamar University, Yemen.
*Author to whom correspondence should be addressed.
Abstract
Due to the huge amount of computer data stored in databases, one centralized database cannot support and provide good performance and availability when contains huge data which used by large number of users. Thus, the distributed database is a good technique to overcome this problem by fragmenting the database and allocating the right database fragmentation in the right site. Many researches present static optimized algorithms of distributed database fragmentation, allocation and replication (Horizontal/ Vertical) at the initial stage of the distributed database design using different or similar techniques, which affect the performance of database system. Therefore, this study aims at reviewing and comparing the best-presented algorithms from the design perspective, with the aim of identifying the strength and weakness points of each algorithm. Furthermore, this study could be considered as the first study that attempts to identify the most critical criteria that were used for comparing the optimized algorithms that have been proposed and used in distributed database fragmentation and allocation.
Keywords: Distributed database, vertical partitioning, horizontal fragmentation, attribute usage matrix, frequency matrix, attribute affinity matrix, Crud matrix