Telecommunication networks are developed over many years (over 100 in the case of the UK) and buried underground. Creating and maintaining accurate inventories of the materials used in their construction is challenging, when these systems were implemented electronic records did not exist and large scale surveys are uneconomic and unreliable. A Bayesian model of the network structure is developed and it is shown how this can be used in combination with the telemetry derived from a DSL network to create estimates that repair the inventory records. This approach was evaluated using a physical investigation of a sample of network elements and these results are presented along with a discussion of the limitations of the approach that have been identified. This result was derived vs an inventory system of approximately 7 million entries and required data processed from approximately 130 Tb of telemetry an applied to over 2 million cable records. This technique is now used to deliver broadband services in the UK.
CITATION STYLE
Ferguson, A., & Thompson, S. (2017). Cable belief networks (best application paper). In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 10630 LNAI, pp. 177–189). Springer Verlag. https://doi.org/10.1007/978-3-319-71078-5_17
Mendeley helps you to discover research relevant for your work.