High Fidelity Data Reduction for Big Data Security Dependency Analyses

Explore a conference talk from CCS 2016 focusing on high fidelity data reduction techniques for big data security dependency analyses. Delve into the challenges and opportunities presented by big data in security contexts, drawing from real-world experiences. Examine the problem statement and design intuitions, including concepts like "expendable" dependencies and controlled dependency loss. Learn how domain knowledge contributes to effective data reduction strategies. Evaluate the proposed approach by comparing it to naive methods, assessing its ability to back-track real attacks, and analyzing resource consumption. Conclude with insights into future work in this critical area of cybersecurity research.

Intro
Background
Opportunities and Challenges
Big Data, Why?
Real World Experiences
Problem Statement
Design Intuitions
"Expendable" Dependencies
Controlled Dependency Loss
Domain knowledge Helps
Design Summary
Evaluation Target
Data Reduction Capability
Compare with Nalive
Back-tracking Real Attacks
Resource Consumptions
Conclusion & Future Work