The purpose of autonomicmanagement is to handle heterogeneity and complexity of the systems. The fundamental principle of autonomic management is to enable self-properties (self-organization, self-configuration, self-optimizing, self-healing) in the devices aiming self-management and a low degree of direct human intervention. Thus, autonomic systems can obtain a certain level of flexibility and adapt themselves to new contexts, user needs or environmental changes. The standard definition for an autonomic systembased on IBM’s autonomic framework includes two entities: managed resource and autonomic manager.

The managed resource is the end system and comprises sensors and effectors. Sensors are responsible for collecting data from a managed resource, as well as the environment while effectors are responsible for sending commands to the managed device. Thus, managers can monitor the environment through the sensors and execute that action through effectors. The knowledge base consists of policies, action plans, among other items.

The autonomic manager provides a control loop called MAPE-K: monitoring, analysis, processing, execution and knowledge base. The monitor function is responsible for collecting data from sensors. The analyze function allows the autonomic manager to analyze the data, comparing this with historical and current data, rules and beliefs, and perform diagnosis. The plan function provides a guide to necessary actions with the help of policies, to achieve goals and objectives. Finally, the execute function controls the execution of the pre-defined.