Autonomic Computing

from Wikipedia, the free encyclopedia

Autonomic Computing is a paradigm used primarily by IBM as a term since the beginning of the 21st century , which includes possibilities of self-configuration, self-optimization and self-healing, especially from a business perspective (and thus also cost aspects).

Nevertheless, the approach of autonomic computing should not be confused with the more demanding autonomous computing or systems of artificial intelligence or robotics or the corresponding investigation of self-organization in cognitive science .

The basic idea of ​​autonomic computing is not new - error detection and error tolerance, for example, are basic performance characteristics of (sub) computer systems. Autonomy in itself is even one of the central techniques of software architecture . The further development of autonomy as a boundary between systems takes place, for example, in service-oriented architectures . The local interpretation of the concept of autonomy is weaker, because it does not refer to self-organizing systems .

What is new about autonomic computing is the requirement and the engineering perspective of self-organization and the level of operation. The latter takes place from the overall perspective of the company or from overall and not just subsystems.

Self-management in this sense is made up of four principles, which are also referred to as self-CHOP :

  • Self- configure (“self-configuration”), automatic configuration of components
  • Self- heal ( "self-healing"), automatic detection and correction of faults
  • Self- optimize ("self-improvement"), automatic observation and control of resources in order to enable them to function as best as possible with regard to the defined requirements
  • Self- protect , predictive detection of attacks and protection against attacks

Levels of maturity of autonomy

Accordingly, there are five levels of corporate IT maturity :

  • Fundamental , in which the individual components of the IT infrastructure are maintained and operated separately.
  • Directed where system management tools are used to collect information centrally.
  • Predictive , whereby possible scenarios can be calculated in advance with the help of analysis methods and tools ("what-if").
  • Adaptive , where computer systems can start automated actions based on information systems and extracted "knowledge".
  • Autonomous , is an IT infrastructure that is completely driven by requirement descriptions and defined goals.

See also

Web links