The Problem of Self-Tuning Databases Has Been Veryimportant For the Last 10 Years. Commercial Database Systems Have Been Addingfeatures For a Long Time and Have Now Reached a Level of Complexity That Makesthem a Difficult Choice As a Central Platform For Information Services Ore-Commerce. It Is Critical That Database Systems Be Easy to Manage, Predictablein Their Performance Characteristics And, Ultimately, Self-Tuning. Self-Tuningdatabase Technology Is Bound to Build on Mathematical Models and to Provideproper Engineering into System Components.Today’S Dbmss Should Be Knob-Free Andmust Be Able to Adapt As Conditions Change, Without Human Intervention. Anautonomic Dbms Should Be Capable of Managing and Maintaining Itself As Well.This Paper Examines the Characteristics That a Dbms Should Have In Order to Beconsidered Autonomic.Optimization of Self-Tuning Models Strives to Minimize Aprogrammable Goal Function That Reflects the System Designer’S Preferences Andthe System Behavior.