Modifikasi Metode Timestamped Change Data Capture pada RDBMS (Relational Database Management System)

Abstract

A normalization process must be applied in the data structure in database designing process in RDBMS, in order to eliminate data anomaly and redundancy. This procedure increases access performance as well as reduces unnecessary storage space. It transforms a single entity into multiple entities, normalization process is often so-called decomposition. A problem occurs, however, after a normalized database is being implemented. When a new data recorded there is nothing wrong with its retrieval, while the retrieval process goes into the last data updated in the case of updating activity. Moreover, when a deletion occurred, the data can never be retrieved. Timestamped CDC method offers a way of tracking data change. Basically, timestamped CDC uses historical record to include timestamp everytime change occurs. Tracking updated record is simply done by the previous research of timestamped CDC, while keeping track on deleted data is impossible since the data vanished within the deletion process. This paper modifies the timestamped CDC method so that data deletion tracking can be achieved as well as data updating. Furthermore, this research also makes several testing procedures to find out performance of the modified timestamped CDC structures.