MOFA: a model for fault and availability in complex services

Asset allocation to optimize reliability and availability (R&A) has been studied in the literature. However, these concepts have not been investigated in combination with degraded performance in realistic reliability-models. Some reliability models, techniques, and tools allow modeling of degradable performance. However, they are mainly used in evaluating R&A of given systems; they do not consider the inverse problem of determining the system parameters that would maximize the R&A. This paper provides a realistic reliability model for complex services; the model can be used to study the asset allocation problem in degradable complex services. This paper uses a simple model to describe the failure characteristics of the system parts. This model provides: (1) an infrastructure for defining R&A in a flexible way, and (2) a way of defining at what level of availability it is worth offering the service, and at what level of failure of resources the system should be declared faulty. This model is then used to 'allocate optimally the available assets' to 'implement a service that is tailored to the availability needs of the user and the situation'. The model is applied to a realistic example of a complex service, viz. a multimedia communication service.