Enhancing covert communications with multiple colluding receivers

Traditional (single receiver) system setup: ? Choose exploit field (e.g. last byte of TCP Timestamp) ? Alice: probabilistically inject parts of coded message into field ? Bob: extract symbols from field, decode to correct errors ?Warden: assume full knowledge of system and keys Can we create undetectable system? Previous detection work: ? Signatures - published exploits thwart easily ? Anomaly - qualitative arguments until statistical methods in [1] ? Brute-Force - never mentioned in literature, significant oversight Thwarting Brute-Force Detection: ? Propose multiple colluding receiver design ? Verify possibility of brute-force in single receiver system ? Show our design's resilience to threat Thwarting Anomaly Detection: ? Propose better quantification technique ? Provide fast approximation