Fig. 1.1) by means of system software, so that in the end no “weak” component of
the system reduces the reliability (right box of Fig. 1.1).
Fault tolerance as a new feature of a system requires additional overhead in both
software and hardware. This area of research became visible in the early 60s [3].
Several attempts and efforts were done to explore ways how to design and classify
fault-tolerant systems. Classifications of measures that implement hardware fault
tolerance were presented in [4–6]. But even after these significant contributions a
general question has not been answered:
Do we know enough to define the system behavior and predict its behavior from our
classification in presence of faults?
In other words, fault tolerance, as a theory might be valued using following
statement:
The power of any theory is in predictions and our ability to apply them.
The usefulness of the predictions depends here on timing, relevance, and
practicality. Structural analysis of variants how to implement system software
support of hardware deficiencies uses the classification of redundancy types and
schemes of its application to achieve fault tolerance. The classification developed
Fig. 3.3 Maximum
reliability of RT FT system
3.2 Connection Between Reliability and Fault Tolerance
15
the system reduces the reliability (right box of Fig. 1.1).
Fault tolerance as a new feature of a system requires additional overhead in both
software and hardware. This area of research became visible in the early 60s [3].
Several attempts and efforts were done to explore ways how to design and classify
fault-tolerant systems. Classifications of measures that implement hardware fault
tolerance were presented in [4–6]. But even after these significant contributions a
general question has not been answered:
Do we know enough to define the system behavior and predict its behavior from our
classification in presence of faults?
In other words, fault tolerance, as a theory might be valued using following
statement:
The power of any theory is in predictions and our ability to apply them.
The usefulness of the predictions depends here on timing, relevance, and
practicality. Structural analysis of variants how to implement system software
support of hardware deficiencies uses the classification of redundancy types and
schemes of its application to achieve fault tolerance. The classification developed
Fig. 3.3 Maximum
reliability of RT FT system
3.2 Connection Between Reliability and Fault Tolerance
15