Privacy analysis of user association log in an enterprise wireless network Keren Tan
Background To protect privacy, we sanitize network trace before sharing them General sanitization mainly focus on randomizing or truncating explicit user identity information * randomize truncate
Motivation Besides those explicit identity information in each log entry, much information that can be linked to a specific user (or a small subset of users) is also implicitly contained in collected traces.
Which one is Snoopy? Person 1 Person 2 Sudikoff, 45min Sudikoff, 55min Bake, 50min Bake, 30min Bake, 60min Gym, 100min Gym, 55min Gym, 80min Sudikoff, 25min NP, 10min Sanitized Person 1 looks like more similar than Person 2?
Challenges Scalability! Duration 3 months Dataset size > 50GB Number of users > Number of APs > 1300