Dissimilar redundancy

Make wise use of dissimilar redundancy (e.g. electrical systems backed-up with a mechanical alternative). The virtue of dissimilar redundancy is that, because the channels are fundamentally different in their design, it is much less likely for an external event to affect them all in the same way (Lloyd and Tye, 1982, p. 94). 

Main analysis for the architectural metrics is to make the safety relevant signal chains transparent and add adequate safety mechanism in case of weaknesses. Inspired by Robert Lusser, the signal chains are a chain of elements and the weakest parts should be enforced by means of safety mechanism. A typical safely mechanism consist of a part that can detect, malfunctions such as fault, errors or failure and a part that could control the malfunctions. It should be able to degrade the system to a safe state or switch to dissimilar redundant functions, which are identified as error free during runtime. Therefore, the entire signal chains and its elements (chain links) need to be identified. The quantification after Erich Pieruschka is primarily used to make the strengths of the chain links comparable. What is important The safety relevant function is first subject of the analysis. The correct functioning of the safety relevant function has to be assured. If this is provided by adequate measures such as implemented safety mechanism and control measures, this forms architecture to safety architecture. 

Regarding dissimilar redundancy dissimilarity can be both in hardware and software. Software dissimilarity is achieved by producing two separate software lequirements/solutions and by the use of two separate teams. Increased workload is countered by being able to limit the amount of testing hy virtue of the replication of computation. The two software lines run asynchronously in two processes and their outputs are added or compared to ensure that no demand is made incorrectly. This achieves high integrity hut at the expense of avaUahUity, so that where passivity cannot be tolerated (e.g. fly-by-wire control systems) such architectures must have an alternative central lane in the event of a failure. 

In view of the conflict between the reliability and the cost of adding more hardware, it is sensible to attempt to use the dissimilar measured values together to cross check each other, rather than replicating each hardware individually. This is the concept of analytical i.e. functional) redundancy which uses redundant analytical (or functional) relationships between various measured variables of the monitored process e.g., inputs/outputs, out-puts/outputs and inputs/inputs). Figure 3 illustrates the hardware and analytical redundancy concepts. 

Given a table of redundant properties, one could calculate dissimilarities as Euclidean distances and use MDS instead of PCA. Whereas the results would be similar, this is usually wasteful, because the number of molecules is typically much greater than the number of properties. However, if the similarities are best calculated from a nonlinear function of the properties, such as Tanimoto coefficients computed from two bit strings, the results would not be similar and nonlinear MDS should be used. One then gets back a set of latent properties (dimensions) for which Euclidean distance approximates the desired similarities. Thus, a simple rule of thumb is For redundant properties as input use PCA, for metric similarities as input use classical MDS, for nonmetric similarities use numerically refined MDS. 

The filter-based methods operate in isolation for ranking the features and do not consider the correlation among the features. Thus, the redundancy among the selected features is not used. To overcome this problem, MRMR method have been used that takes into account both minimum redundancy and maximum relevance criteria to select the additional features that are maximally dissimilar to the already identified features. 

In multiple-version dissimilar software, a set of two or more computer programs are developed separately (and independently) to satisfy the same functional requirements. Errors specific to one of the versions should be detected by comparison of the outputs between the different versions. When multiple-version dissimilar software is used in redundant computer systems, the likelihood of the same errors in both systems is theoretically significantly reduced, thereby achieving a higher level of safety for the system. Multiple-version dissimilar software involves applying the dissimilar design concept to software. 

N-version programming is a system design technique that involves producing two or more software components that provide the same function, but in different ways. This approach is intended to avoid sources of common errors between redundant components. N-version programming is also referred to as multi-version software, dissimilar software, or software diversity. 

In terms of availabihty, FBW architecture is based on a high level of redundancy. The introduction of cormnon hardware resources of integrated modular avionics induces the de facto piesenee of common elements in the system. In the knowledge that regulation insists that a simple failure should not have catastrophic effects, it would be necessary for eertain system functions (or sub-functions such as of ultimate rescue type) not to use these hardware resources or that the integrated modular avionics resources provide dissimilar platforms. The concept of decentralizing the electronics control from the actuators (section 6.3.1) to bring the... 

The safety objectives for the FEW part of the system (PRIM plus SEC) have been defined without benefiting from this ultimate back-up. However, on the A400M, the back-up is fidly capable of flight and safe landing, and this must be demonstrated for mihtaiy qualification. This results in a decreased need for dissimilarity between PRIM and SEC they share a common type of hardware (but different functions and software). Tolerance to a design or hardware manufacturing error is ensmed by the functional dissimilarity between PRIM and SEC on the one hand, and the total hardware dissimilarity between PRIM/SEC and the ultimate rescue redundancy, on the other. 

Fu, Y.B., 2006. Redundancy and distinctness in flax germplasm as revealed by RAPD dissimilarity. Plant Genet. Resour. 4, 117-124. 

---