A framework for manipulating vacuumed data in temporal relational database

The Temporal database is one of the databases that manipulate by append-only policy instead of updating in-place. The data in these databases have two main features: valid-time and transaction-time. Since, the data aren´t deleted in temporal database; instead they are increasingly expanded and grown up, it´s necessary to adopt a mechanism for controlling the volume and capacity of the database. In such a database a large quantity of the information are fetched less, while some are fetched more, so that it is essential to use a vacuuming data method as well as physical deletion technique to control the database volume. In the present research, we introduce an intelligent vacuuming system based on an unintelligent model of SDVMT which attempts to vacuum the data based on the extent of data importance, transaction time and valid time using a distributed middleware platform. The intelligent model increased the accuracy of the unintelligent model. This model behaves intelligently by learning from the behavior of the system administrator, end user and the server´s performance. Therefore, the importance of data is identified by analyzing the behavior of end users. In such a process, the servers are classified based on their performance by continuous monitoring of servers and observing the behavior of system administrators in data vacuuming.