CYBER 2018 : The Third International Conference on Cyber-Technologies and Cyber-Systems Prototype Open-Source Software Stack for the Reduction of False Positives and Negatives in the Detection of Cyber Indicators of Compromise and Attack Hybridized Log Analysis Correlation Engine and Container-Orchestration System Supplemented by Ensemble Method Voting Algorithms for Enhanced Event Correlation Steve Chan Decision Engineering Analysis Laboratory San Diego, California U.S.A. email:
[email protected] Abstract—A prototypical solution stack (Solution Stack #1) black boxed commercial offering itself. Accordingly, with chosen Open-Source Software (OSS) components for an MSSPs are endeavoring to put forth their own offerings experiment was enhanced by hybridized OSS amalgams (e.g., (also for market differentiation), in a white box fashion; for Suricata and Sagan; Kubernetes, Nomad, Cloudify and Helios; the purposes of adhering to Joyce’s recommendation, the MineMeld and Hector) and supplemented by select modified white box approach can be, in some cases, more effective algorithms (e.g., modified N-Input Voting Algorithm [NIVA] modules and a modified Fault Tolerant Averaging Algorithm than the black box approach, but it is a much more difficult [FTAA] module) leveraged by ensemble method machine pathway in terms of the sophistication needed to understand learning. The preliminary results of the prototype solution and appropriately orchestrate the various involved stack (Stack #2) indicate a reduction, with regards to cyber subsystems. By way of example, a Verizon Data Breach Indicators of Compromise (IOC) and indicators of attack Report had articulated that those with robust log analysis (IOA), of false positives by approximately 15% and false and correlation were least likely to be a cyber victim; yet, negatives by approximately 47%.