DeepL Replay - Time series replaying of semi structured machine generated data

Abstract

OI is vital for an organization’s decision making process at the strategic level. Machine generated data is the richest source which could be utilized to gain OI and is available and applicable for any domain. Many commercial products exists which utilizes machine generated data to produce OI. Multiple researches are also conducted on utilizing machine generated data to produce OI. However these researches and commercial products do not address the following problems: Replaying of machine generated data in order to re-create a problem that is hard to re-create. Correlating multiple sources of machine generated data to produce sensible information. Further elaborating on the problem, commercial products which exists to produce OI use unstructured data in order to do so and therefore has failed to make sense of the data but rather merely produce statistics based on regular expressions. Research has been carried out in the domain of structured log analysis to overcome this problem. However a change to the structure of the log file will cause the process to fail to make sense of the data if not adapted manually according to the new structure of the log file. The research will elaborate on using semi-structured machine generated data to produce OI to overcome the problem of producing OI using unstructured and structured data. Therefore the purpose of the project is to research, design, develop, test and evaluate a prototype which is capable of replaying semi-structured machine generated data based on the time-series information.