As communications networks are expanding to larger areas the control and maintenance of routing information are becoming a formidable task. To cope with its size and complexity and to make the network reliable and scalable hierarchical network has been proposed with new features to support the information infrastructure. However, the network hierarchy adds more complications to the network design and implementations and that hampers the network reliability and quality of services. Conformance testing is known to be a powerful tool for network fault detection yet most of the works in the published literature are on networks without hierarchy. We present probabilistic algorithms for testing hierarchical networks along with the added features. Based on a formal model of the networks, we provide a formal analysis that shows that our probabilistic algorithms guarantee a high fault coverage with a feasible number of tests. To further reduce the number of tests we identify test equivalence classes and that enables us to significantly reduce the number of tests yet without losing the fault coverage. Experimental results on Internet OSPF protocol are reported.
展开▼
