Data Exfiltration via Web/Cloud

Description of Use Case

The scoring for this use case overwhelmingly favored specialized security tools that can inspect the content of data and understand cloud-native applications. General-purpose network and OS logs, while providing broad coverage, consistently scored lower due to a lack of deep, contextual insight. The highest scores were achieved by sources that could answer “What data left?” not just “Did data leave?”.

Techniques Evaluated

  • T1567 Exfiltration Over Web Service

  • T1041 Exfiltration Over C2 Channel

  • T1071 Application Layer Protocol

  • T1132.001 Data Encoding: Standard Encoding

  • T1102.001 Web Services: Bidirectional Communications

  • T1020 Automated Exfiltration

  • T1048 Exfiltration Over Alternative Protocol

  • T1056 Archive Collected Data

  • T1001 Data Obfuscation

  • T1537 Transfer Data to Cloud Account

Top Scoring Log Sources

Log Source

Score

DLP (Data Loss Prevention) Alerts

28.7

EDR: Data Exfil Heuristics

27.1

CASB (Cloud Access Security Broker) Logs

26.4

EDR: Suspicious Child Process

26.7

Sysmon: EID 8: CreateRemoteThread

26.0

EDR / Endpoint Telemetry

25.8

AWS: GuardDuty Findings

25.8

Azure: Sentinel Detections

25.8

GCP: Security Command Center Findings

25.8

EDR: Remote Access Software Detections

25.3

Evaluation of Log Source Types

Content-aware alerts (DLP, CASB) were the undisputed champions. Network traffic metadata (NetFlow, Firewall logs) ranked significantly lower, establishing a clear pattern: seeing the content and destination context is more valuable than seeing the volume and connection data alone. Process events were middle-tier: useful for seeing a tool like 7z.exe run, but not for seeing the data leave.

Technology Comparison: Specialized Security Tools (DLP, CASB, EDR) >> Cloud-Native Security >> Native OS / Network Gear

Scoring Data

The raw TC scores broken down by metric can be found here: Data Exfiltration via Web or Cloud Services