Model Federation and Probabilistic Analysis for Advanced OSS and BSS

Abstract

Advanced OSS and BSS will be expected to operate cooperatively and across multiple domains and business layers. This can be reached with shared information models providing a comprehensive insight into the entire operated heterogeneous environment. This paper contributes to this vision in two respects. It first introduces a technique for creating a federated information model by inter-relating existing and potentially very different domain specific models. Furthermore, the resulting federated model is used as structural base for defining probabilistic analysis with a Bayesian network. This demonstrates how valuable insights can be obtained through model federation rather than solely relying on separated models reaching only a limited set of information.