Key points are not available for this paper at this time.
We benchmark the implementation-level overhead of ETSI GS QKD 004 and ETSI GS QKD 014 application-to-KMS key-delivery interfaces in NS-3/QKDNetSim, isolating interface semantics as the only experimental variable within the same KMS reference implementation. We quantify interface-dependent effects by measuring (i) effective key-material debiting at the KMS buffers, (ii) application–KMS message/byte exchange, and (iii) peer KMS coordination load, under identical link, traffic, and cryptographic settings. The results show how session-based key streaming (ETSI 004) versus request/response discrete key delivery with key identifiers (ETSI 014) reshapes signaling overhead, KMS load, and effective utilization in both single-link and relay (multi-hop) settings. All results are obtained from reproducible simulations in NS-3/QKDNetSim; no measurements from deployed QKD equipment are used.
Lauterbach et al. (2026) studied this question.