In some circumstances, it is desirable to ensure the system continues to operate even if there is an internal failure. An aircraft navigation system should be able to operate even if an internal dc-dc regulator fails, for example.
Not everything within some systems benefits by being fault tolerant.
For example, a failure of a cabin reading light over a passenger seat is not critical to the safe operation of the aircraft, thus is likely not created to be fault tolerant. One criterion to determine what should be fault tolerant is the criticality of the function the system provides.
This also applies to specific subsystems within a system allowing some elements to be created fault tolerant and others within the system not.
How to Determine What to Should be Fault Tolerant
Besides considering the function, the elements (components) of a system each contribute to the overall functional performance.
Evaluating each element using a few questions may help determine the specific elements that may benefit with fault tolerance.
How critical is the component to the essential function? If the component creates a single point failure for the essential function it is a candidate.
How likely is the component to fail?
This becomes a tradeoff considering the added complexity of a fault tolerant design versus the need to employ the alternate means to provide the function if an element fails. Complexity, weight, and cost are just a few of the factors to consider. If an element is very unlikely to fail, the effort to create a fault tolerant system may be wasting resources better used to improve other elements of the system.
How expensive are the fault tolerant options? Also, part of the tradeoff, the more expensive and complex the fault tolerant designs both the criticality and the likelihood of failure must likewise be high.
Some systems fault tolerance is obvious, such as 911 emergency services phone systems, some banking and commerce system, safety critical systems of aircraft, public transit, or chemical plants. And sometimes the decision to add some level of fault tolerance is not so clear and requires careful consideration of the costs, benefits, and criticality.
Related:
The Downside of a Fault Tolerant System (article)
Fault Tolerance Basics (article)
Benefits of Fault Tree Analysis (article)
Leave a Reply