Optimal eavesdropping on QKD without quantum memory

TL;DR

This paper analyzes the security of several quantum key distribution protocols when the eavesdropper lacks quantum memory, providing new bounds and showing increased tolerable error rates under this assumption.

Contribution

It introduces a method to bound security of QKD protocols against memoryless eavesdroppers, improving previous limits on tolerable error rates.

Findings

Maximum QBER for secure key extraction in six-state protocol increased from 12.6% to 20.4%.

Provides new security bounds for six-state and SARG04 protocols against memoryless adversaries.

Demonstrates that lack of quantum memory in eavesdroppers significantly enhances protocol security.

Abstract

We consider the security of the BB84, six-state and SARG04 quantum key distribution protocols when the eavesdropper doesn't have access to a quantum memory. In this case, Eve's most general strategy is to measure her ancilla with an appropriate POVM designed to take advantage of the post-measurement information that will be released during the sifting phase of the protocol. After an optimization on all the parameters accessible to Eve, our method provides us with new bounds for the security of six-state and SARG04 against a memoryless adversary. In particular, for the six-state protocol we show that the maximum QBER for which a secure key can be extracted is increased from 12.6% (for collective attacks) to 20.4% with the memoryless assumption.