Alojzij Blatnik (2012) Forecasting critical events in provision of computational resources. EngD thesis.
Abstract
Information technology companies usually own many servers, network devices, printers, uninterruptible power supplies, etc. A large number of devices is difficult to manage and to assure their uninterrupted service and early problem detection a periodic device monitoring is essential. We analyse the collected data and try to detect anomalies. We describe methods for automatic or semi-automatic detection of anomalies in the functioning of selected devices as well as adequate settings of their parametres.
Actions (login required)