QKD: a million signal task

TL;DR

This paper reviews the security bounds in quantum key distribution for finite key lengths, emphasizing that secret keys cannot be securely generated with fewer than 10^5-10^6 signals, based on fundamental estimates.

Contribution

It provides a simplified derivation of the minimum number of signals needed for secure quantum key distribution, connecting detailed protocol studies to basic estimates.

Findings

Secret key extraction requires at least 10^5-10^6 signals per run.

Security bounds can be derived from fundamental estimates.

Practical implementations must process a large number of signals for security.

Abstract

I review the ideas and main results in the derivation of security bounds in quantum key distribution for keys of finite length. In particular, all the detailed studies on specific protocols and implementations indicate that no secret key can be extracted if the number of processed signals per run is smaller than 10^5-10^6. I show how these numbers can be recovered from very basic estimates.