Developing an Efficient Failure Prediction Model in Support of a Distributed Operating System for Autonomic Networks