The document discusses AT&T's use of a Hadoop-based approach for near real-time outlier detection and interpretation of cyber threats. It provides context on the challenges of detecting threats at AT&T's network scale and data volume. The presentation outlines AT&T's history of threat capabilities, the need for a big data solution, and their transition to a Hadoop-based threat analytics platform for ingesting and analyzing over 5 billion daily network events to detect outliers and threats.