Towards the unconditional security proof for the Coherent-One-Way protocol Mhlambululi Mafu, Adriana Marais, Francesco Petruccione Quantum Research Group, School of Physics and National Institute for Theoretical Physics, University of KwaZulu-Natal E-mail: mhlambululi.mafu@gmail.com, adrianamarais@gmail.com and petruccione@ukzn.ac.za Abstract. Quantum Cryptography, one aspect of which is Quantum Key Distribution, provides the only physically secure and proven method for the transmission of a secret key between two distant parties, Alice and Bob. The goal of QKD is to guarantee that a possible eavesdropper (Eve), with access to the communication channel is unable to obtain useful information about the message. The Coherent-One-Way protocol is one of the most recent practical QKD protocols. However, its unconditional security proof still remains unrealized. We therefore present a necessary condition for the security of the COW protocol, and show that Bob’s measurements are described by non-commuting POVM elements which satisfies this condition. 1. Introduction The Coherent-One-Way (COW) Quantum Key Distribution (QKD) protocol was first proposed by Gisin et al. [1] and belongs to a class of the so-called distributed-phase-reference protocols [2]. Currently no lower bound is known for the unconditional security of this protocol. The existing tools for proving security of protocols against the most general attacks fail to apply to this protocol in a straight forward way. While security proofs for some limited attacks exist, the unconditional security proofs still remain unrealized. This paper presents an improvement in this direction for the COW QKD protocol by presenting a necessary security condition for the most general kind of attacks which is a step towards the unconditional security proof. Various attacks have been studied for the COW QKD protocol. Security against intercept and resend attacks based on unambiguous state discrimination has been shown [4] as well as for general individual attacks [3]. Upper bounds for the error rates for the security of COW protocol in the presence of large collective attacks as well as collective beam splitting attacks [3] have also been derived [2]. Security against sequential attacks [5] based on unambiguous state discrimination have also been shown [4]. The security against the most general attacks is still elusive because the present tools for proving security of protocols in general cannot be adopted in a straightforward way. In spite of these proofs against limited examples of kinds of attacks, it still remains unclear how the unconditional security proof can be realized. This is mainly because this class of protocols use coherent sequences of signals which are not symmetric as opposed to qubits in other classes of protocols. Again, these protocols move away from the symbol-per-symbol type of coding [7]. The notation and the formalism to be used to develop a full unconditional security