| The data from the four autonomic components are available as shared knowledge. This shared knowledge comprises many types of data, such as metrics, commands, system events, logs, performance data, and policies. It can be described as follows: -
The sensors and effectors collect the information from an element, such as a file, database, or a similar system segment. -
The data from the sensors and effectors is analyzed during the analysis phase to determine if the policy is being adhered to, or if corrective action is needed, an example of self-managing. -
A policy is defined as a set of behavioral constraints or preferences that influence the autonomic manager. -
The plan component is responsible for interpreting and translating the polices for the managed element. -
The results, metrics, and commands are stored as knowledge. This is the basic structure of the control loop. It relies on the technique of feedback control optimization. Control loops will be required to work with many different types of computing technologies and different software that exist in the IT world today. As the technology of control loops develops, more intelligence can be built in. There may be thousands of control loops in big, complex IT systems, and they will be required to work together. This is referred to as autonomic manager collaboration. |
