Big data for digital forensics
Tutor / director / evaluatorHernández Serrano, Juan
Document typeMaster thesis
Rights accessOpen Access
Digital Forensics and its sub-branch Network Forensics are important and relevant topics which have gained further attention with the DDoS attacks delivered by botnets. This work focuses on a novel IDS solution called: SLIPS. This is a free software that uses Machine Learning to detect malicious behaviors in a network with the use of Markov Chain based detection and previously trained models. A major limitation of SLIPS lies on its performance, and this work also touches on the topic of Big Data, and more specifically MapReduce, in order to aid SLIPS with a better resource utilization. With the redistribution of SLIPS tasks across workers, adding a pre-processing of data, the proposed solution using MapReduce presented performance improvements of up to 433 times with the datasets tested.