In order for any system to successfully operate, it must achieve “dependability.” For instance, in the computing technology world, “dependability” is defined as a system’s “reliability, availability, safety, security, survivability and maintainability.”18 In simpler terms, it means reliance can “be placed on the service [a system] delivers.”19 Manual decision making and implementation in a system managing tens or even hundreds of demands is possible. But as a system becomes increasingly complex, successful manual management becomes a less likely outcome.
Using the computing technology example, consider that for a system to function, hardware components require software to tell them which tasks they must perform. Over time, both software (programming) and hardware (system components) have become so complex that it is impossible to manually handle constantly changing and often conflicting demands in a timely and decisive way. This problem in computing drove International Business Machines (IBM) to coin the phrase “autonomic computing” in 2001, which described the vision for developing self-managing computing systems.20
Denoting “self-management,” an autonomic system works through an “automation of responsibility including some decision making for the successful operation of the system.” 21 Unlike an “autonomous” system, an “autonomic” system is classified as ‘involuntary,’ and ‘spontaneous.’ An autonomic system operates in a dynamic environment and is effectively able to help reduce burdens on a complex system through the automation of lower level, but still vital, decision making and implementation.22
Although “autonomic computing” is derived from the biological autonomic nervous system, there is a key difference between the biological and computing versions: autonomicity must be constructed in computing, it does not naturally occur as it would in a healthy organism. In order to construct autonomic functions, four self-managing objectives must first be realized:  Self-configuring,  Self-healing,  Self-optimizing and  Self-protecting (see Table 3).23 These four objectives cannot operate independently, but instead must be realized as parts of a whole system.
Within each of these four self-managing objectives (the “Four Self-”s) is another control loop. It describes how each of the “Four Self-”s objectives is realized individually. The control loop is abbreviated as “MAPE-K” and stands for Monitor, Analyze, Plan, Execute- Knowledge. It describes a process of continual adjustment according to a constantly changing internal and external environment.
This process can be described as the interaction of two channel types: a Sensory Channel and a Motor Channel. The Sensory Channel describes the capacity to sense the state and changes in its internal and external environment (or “Monitoring” and “Analyzing” the current state and environment). The Motor Channel describes how an attribute, reacting to and countering the effects of changes, adapts to maintain equilibrium (or the “Planning” and “Execution” of decisions and change).24 The successful operation of this process means that a system is constantly adapting in order to prevent faults and to handle faults when they inevitably occur. The end goal is not a system without faults. The goal is a system capable of ensuring that faults do not cause the entire system to collapse.