Privacy-Preserving Sharing and Correlation of Security Alerts

We present a practical scheme for Internet-scale collaborative analysis of information security threats which provides strong privacy guarantees to contributors of alerts. Wide-area analysis centers are proving a valuable early warning service against worms, viruses, and other malicious activities. At the same time, protecting individual and organizational privacy is no longer optional in today's business climate. We propose a set of data sanitization techniques that enable community alert aggregation and correlation, while maintaining privacy for alert contributors. Our approach is practical, scalable, does not rely on trusted third parties or secure multiparty computation schemes, and does not require sophisticated key management.