Prediction of resource failure uncertainty for ubiquitous applications in grid environment

In the recent years, the use of grid computing has been extended to ubiquitous computing that demands high computing capabilities. The tremendous development of mobile and embedded devices in ubiquitous computing applications and their communication requirements forms a large data set that need to be computed efficiently. However, these devices have limited resource capacity such as processing element (PE), storage, battery power, bandwidth, and so on. Hence it is required to transfer requests from ubiquitous applications to grid computing environment for computation. Nevertheless, the major issue with resource sharing lies in the uncertainties like resource failure, communication link failure, etc., during massive job arrival. Thus, it is important to consider uncertainties in the formulation of a grid scheduling problem. In this paper, we propose a model that demonstrate the uncertainty of resource failure using probabilistic Markov chain and Poisson distribution. These techniques are exploited as maximum likelihood estimator. The implementation shows prediction accuracy of uncertainty.