Self-Optimizing Database Architecture

Abstract

Abstract: Over the most recent decades, analysts and database service providers have fabricated devices to help DBAs (Database Administrators) in various parts of framework tuning and the actual design of the database. Most of this past work, regardless, is fragmented on the grounds that it expects people to come up with an official agreement or judgement about any modifications to the data in the database and fix issues after they happen rather than preventing such cases from taking place or adjusting to these changes automatically. What is required for a really "self-driving" database management system (DBMS) is another way of approaching this that is intended for independent activity and automatic decision making. This is different from prior endeavors since all angles of this framework are constrained by a coordinated arranging part that not just enhance the framework for the current responsibility, but in addition to this, it also predicts future responsibility that might take place and prepares itself for such not-so-common occurrences and adjusts to them as required while keeping the efficiency of the operations as close to normal as possible. With this, the DBMS can uphold all the past tuning procedures without requiring a human to decide the right way and proper opportunity to use them. It likewise empowers new advancements that are significant for current DBMSs (Database Management System), which are impractical today because of the fact that the intricacy of overseeing these frameworks has outperformed the abilities of human specialists who are supposed to tune them and make changes when required. Keywords: Database Management System, Database Administrator, Forecasting, Long Short-Term Memory, Recurrent Neural Networks