Fuzzy reliability modeling of distributed client -server systems

With the proliferation of client-server systems there is a need for new approaches to determine their reliability in the uncertain environment of the internet. The proposed methods analyze client-server reliability by accounting for uncertainties in the operational environment of the internet. The uncertainties are addressed with a fuzzy paradigm applied in a tractable, verifiable reliability model of client-server systems. For computer hardware, fuzzy sets were used to address vague reliability data using bi-state fuzzy probabilistic theories. For software and operating systems, fuzzy sets account for the uncertainties in bi-state failure data using fuzzy possibility theories. Client-server architectures, such as hypercubes, and multistage interconnection networks are examined via logical and numerical approaches. Addressing uncertainties in the communication performance and expectations of client-server transactions requires a possibility-based fuzzy state model. Resulting from these approaches are fuzzy models that can handle uncertainty in failure data and the uncertain state of communication performance of the internet. The effect of fuzzy system reliability on the reliability of fault-tolerant networks was evaluated by fuzzy entropy and shown to be system-dependent. The possibility-based fuzzy Internet state was shown to have a large impact on the reliability of the system, depending upon the performance expectations. In conclusion, major sources of uncertainty in client-server systems are incorporated into a validated model for determining the reliability of client-saver systems placed on the internet