The article presents the development of a concept for the distribution and migration of data to several databases within a distributed knowledge base. Thus a concept for a load-dependent and application-specific migration of data is introduced which takes into consideration the number and sort of accesses to the data. The main part of the migration method is a distribution function, which generates an optimal distribution of data on the databases according to the defined aims. It uses all information of the profile of accesses and evaluates this information with the help of several decision parameters of different priority. A quick reaction on changes in the profile of accesses is guaranteed by a cycle protocol, which consists of three parallelly running phases with a sort of pipeline structure
Published in:
Database and Expert Systems Applications, 1996. Proceedings., Seventh International Workshop on
Date of Conference: 9-10 Sep 1996