Privacy amplification (PA) is an essential part in a quantum key distribution
(QKD) system, distilling a highly secure key from a partially secure string by
public negotiation between two parties. The optimization objectives of privacy
amplification for QKD are large block size, high throughput and low cost. For
the global optimization of these objectives, a novel privacy amplification
algorithm is proposed in this paper by combining multilinear-modular-hashing
and modular arithmetic hashing. This paper proves the security of this hybrid
hashing PA algorithm within the framework of both information theory and
composition security theory. A scheme based on this algorithm is implemented
and evaluated on a CPU platform. The results on a typical CV-QKD system
indicate that the throughput of this scheme (261Mbps@2.6*10^8 input block size)
is twice higher than the best existing scheme (140Mbps@1*10^8 input block
size). Moreover, This scheme is implemented on a mobile CPU platform instead of
a desktop CPU or a server CPU, which means that this algorithm has a better
performance with a much lower cost and power consumption.

By admin