Collusive Attacks to "Circle-Type" Multi-party Quantum Key Agreement Protocols

TL;DR

This paper reveals that existing circle-type multi-party quantum key agreement protocols are vulnerable to collusive attacks, compromising fairness, and introduces classifications of protocols based on their resistance to such attacks.

Contribution

The paper identifies a common loophole in circle-type MQKA protocols and classifies protocols into circle-type, complete-graph-type, and tree-type based on their security against collusive attacks.

Findings

Circle-type MQKA protocols are vulnerable to collusive attacks.

A common loophole severely affects fairness in circle-type protocols.

Complete-graph-type protocols resist collusive attacks, unlike circle-type.

Abstract

We find that existing multi-party quantum key agreement (MQKA) protocols designed for fairness of the key are, in fact, unfair. Our analysis shows that these protocols are sensitive to collusive attacks; that is, dishonest participants can collaborate to predetermine the key without being detected. In fact, the transmission structures of the quantum particles in those unfair MQKA protocols, three of which have already been analyzed, have much in common. We call these unfair MQKA protocols circle-type MQKA protocols. Likewise, the transmission structures of the quantum particles in MQKA protocols that can resist collusive attacks are also similar. We call such protocols complete-graph-type MQKA protocols. A MQKA protocol also exists that can resist the above attacks but is still not fair, and we call it the tree-type MQKA protocol. We first point out a common, easily missed loophole that severely compromises the fairness of present circle-type MQKA protocols. Then we show that two dishonest participants at special positions can totally predetermine the key generated by circle-type MQKA protocols. We anticipate that our observations will contribute to secure and fair MQKA protocols, especially circle-type protocols.