A robust fault modeling and prediction technique for heterogeneous network

This paper proposes a modelling technique for fault identification in a heterogeneous network. Heterogeneous network is the integration of wired network, cellular network, wireless local area network (WLAN), and Adhoc network, etc. There may be several numbers of causes and sub-causes of fault in each type of network. These causes of fault must be mitigated in order to have a secure and reliable system. The paper introduces a model based approach that starts with the enumeration of possible causes of each fault over the heterogeneous network. A framework is then created by such causes which diagrammatically describe the causes and sub-causes responsible for the occurrence of a fault. This paper suggests a conceptual fault cause tree model based on the probabilistic framework for ranking the faults and their possible causes. An effective mitigation strategy is required to mitigate the causes and sub-causes. Once mitigation of the cause creating a fault is achieved, the system is analyzed and tested for accuracy and reliability. The Proposed technique assures that all sub-causes even at the lowest level of abstraction is taken into consideration in making the system more robust against the particular fault. The proposed model is also used to analyze the faulty probability of a heterogeneous network.