7 Tuning statistical logic

 

This chapter covers

  • Tuning statistical constructs to create better security analysis capabilities
  • Uncovering malicious beaconing that uses unexpected communication channels
  • Capturing packets to gain visibility into a threat execution

In this chapter, you will learn and practice building and using more involving statistical constructs in your threat-hunting expeditions. I want you to experience different approaches, different techniques, and different tools in the statistical toolkit to uncover threats.

First, I introduce approaches such as random time jitter and beaconing to demonstrate that using only standard deviation (chapter 6) is insufficient to uncover threats. I also introduce statistical techniques such as quantiles, which can enhance analytical capabilities to uncover anomalies. Next, I describe how to use density distribution functions to detect data exfiltration, anomalies, and outliers.

You may want to play the soundtrack to The Empire Strikes Back (Star Wars: Episode V) in the background while we go through the first scenario: uncovering beaconing with random jitter. Why? All will be revealed as we proceed.

7.1 Beaconing with random jitter

7.1.1 Relying on standard deviation only

7.1.2 Enhancing the analytic techniques with interquartile range

7.1.3 Interrogating the first suspect

7.1.4 Avoiding confirmation bias

7.1.5 Analyzing the data further

7.1.6 Hunting for patterns

7.1.7 Analyzing fields of interest

7.1.8 Interrogating the second suspect

7.2 Exercises

7.3 Answers to exercises

Summary