▼ The purpose of this thesis is to investigate availability distributions and the probability of Service Level Agreement (SLA) violation for telecommunication services requiring high availability, such as 0.99999. This research deals with availability SLAs between information technology service providers, such as carriers, and user corporations. A combination of five different reliability/maintainability scenarios and three different time intervals have been used to generate availability distributions in order to assess the efficacy of availability guarantees. Moreover, Markov and semi-Markov models are used to provide a comparative analysis of the availability distributions and probability of service level violation. In the semi-Markov model, long tail and short tail lognormal distributions are used for the repair time. A Monte Carlo simulation program is developed in JAVA to achieve the research goal. From the generated distributions it is found that there are significant chances of SLA violation, given a wide range of different reliability/maintainability levels used to achieve 0.99999 availability, and the SLA time interval. Importantly, the results are fairly sensitive to the tail of the repair distribution, meaning it is essential to understand repair distribution in order to assess chances of SLA violation. Moreover, results suggest that the way to dramatically decrease the probability of SLA violation is through availability engineering margin. Delivering availability beyond 0.99999 through redundancy and responding more rapidly to failures greatly diminishes the chances of violating a 0.99999 SLA. Advisors/Committee Members: Snow, Andrew P.

▼ This thesis explores the implementation of a neural network for analysis and decision-making purposes within the realm of network management and serves as a stepping stone toward a solution for a much larger problem. The thesis will discuss the current landscape of network management software and explore the possible benefits of a neural network augmented solution, test the efficacy of a neural network in several scenarios where it must classify network element performance data, and compare these results against more familiar linear regression methods. Results showed that the neural network was far superior at classifying the results. However, the general applicability of these results is not yet known since the dataset used was from a single commercial Internet Service Provider. Advisors/Committee Members: Hoag, John.