Anomaly detection – spotting intrusions in Windows environments
Let’s proceed with the hypothesis we made in Figure 11.5. Let’s say that the infrastructure counts more than 30,000 endpoints and the search was run for the past 7 days. The analyst received a shocking 20 million results. What should their approach be to find any threats?
Well, there are four threat-hunting techniques as defined by ACE Responder that could be applied:
- Aggregation or stacking: https://x.com/ACEResponder/status/1674564539776368643. This involves counting distinct values for one or more fields and analyzing the bounds for outliers.
- Clustering: https://x.com/ACEResponder/status/1675984303568695296. This is the grouping of data based on similar features when the number and nature of different groups may not be fully understood. The analysis can be performed using algorithms suggested (i.e., K-Means, DBSCAN, GMM, K-Modes)..
- Searching: https://x.com/ACEResponder/status...