SNIP_ADAPTIVE_PLATFORM_FIGURE_Figure_7_28

Figure 7.28: General Fault Tolerance scheme. From the perspective of Execution Management , fault activation occures when resulting Function Group State or combination of such is requested. Due to the different nature of faults, these can lead to various types of deviations from expected functional behavior and finally result in erroneous system functionality either in terms of correct computational results or timing response. In general, the implementation of fault tolerance mechanism is based on two consistent steps - Error Detection and subsequent Error Recovery . The major focus of Error Detection during Design Phase activities and thus the focus of Fault Tolerance in this specification is on the analysis of potential Failure Modes and the consequent error detection mechanisms that should later be incorporated into the implementation. In contrast, Error Recovery consists of actions that should be taken in order to restore the system’s state where the system can once again perform correct service delivery . Binding ofError Detection and Recovery Actions should be a subject of platform wide fault tolerance model. Remark:The remainder of this section is the subject for elaboration for the next release of this specification. Provision for fault-tolerance mechanisms will consider possible faults, how they can lead to errors within Execution Management and the mechanisms that are introduced to ensure error detection.