Towards scalable system-level reliability analysis

Abstract

State-of-the-art automatic reliability analyses as used in system-level design approaches mainly rely on Binary Decision Diagrams (BDDs) and, thus, face two serious problems:(1) The BDDs exhaust available memory during their construction and/or (2) the final size of the BDDs is, sometimes up to several orders of magnitude, larger than the available memory. The contribution of this paper is twofold: (1) A partitioning-based early quantification technique is presented that aims to keep the size of the BDDs during construction at minimum. (2) A SAT-assisted simulation approach aims to deliver approximated results when exact analysis techniques fail because the final BDDs exhaust available memory. The ability of both methods to accurately analyze larger and more complex systems than known approaches is demonstrated for various test cases.

Reliable system design using decision diagrams in presence of hard and soft errors

“…7 shows the comparison between the number of nodes needed for pure BDD, the state-of-the-art reliability model as defined in [9], and for SDDs, which can be constructed in both cases. Please note that the number of required nodes represents the memory requirement.…”

“…As described in the previous section, the lifetime reliability value for a design alternative can be calculated using the SDDs constructed using the SSG input to the system design process. This, however, is not true for the state-of-the-art reliable system design as given in [9]. Their technique requires the construction of a new decision diagram for every design alternative that must be evaluated, which is followed by the algorithm to calculate the resulting reliability.…”

“…Our approach was found to be efficient in terms of evaluation time and, in many cases, in terms of memory requirements too. Note that the SEDD does not suffer from the second disadvantage pointed out in [9].…”

“…In [9] Glaß et al have pointed out two disadvantages of their previous approach detailed in [7]. First, for many specification graphs the size of the associated BDD is often up to several orders of magnitude larger than the available memory.…”

“…Secondly, many additional temporary nodes are required during the construction of the final data structure, thus further exacerbating the memory problem. To address mentioned disadvantages Glaß et al [9] have presented two methods. The first one overcomes the second disadvantage only, whereas the second method results in just an approximate calculation of reliability values.…”