DH-EAC: Design of a Dynamic, Hierarchical Entanglement Access Control Protocol

We propose Dynamic, Hierarchical Entanglement Access Control (DH-EAC), a pure-quantum protocol for fair and anonymous allocation of scarce entanglement across wide-area quantum networks composed of many quantum LANs (QLANs). Prior Dicke-state-based pure-quantum MACs resolve contention by local measurements without classical signaling, but they mainly target a single QLAN under static conditions; extending them to wide-area, dynamic settings while avoiding post-selection reconciliation remains open. DH-EAC adopts a two-layer pure-quantum lottery: the outer layer selects winning QLANs and the inner layer selects winning nodes within each winning QLAN. A key design principle is that both the winning set and the per-QLAN quota are fixed by measurements alone, so the contention loop requires no classical round trip. The protocol thus aims to jointly satisfy anonymity (no node IDs revealed until decisions are fixed) and fairness (bias suppression under heterogeneous QLAN sizes). We also provide analytical models for success probability and latency under a standard i.i.d. loss model, and we evaluate DH-EAC against two baselines - single-layer Dicke within one QLAN and a classical GO-driven allocator - using a minimal, reproducible set of scenarios. Metrics include success probability, end-to-end latency, throughput, and Jain’s fairness index. The results indicate that DH-EAC offers an implementable design point in the space of entanglement access control, balancing pure-quantum contention resolution, anonymity, and scalability for multi-QLAN networks.

💡 Research Summary

The paper introduces Dynamic Hierarchical Entanglement Access Control (DH‑EAC), a purely‑quantum protocol designed to allocate scarce entanglement resources across a wide‑area quantum network composed of multiple quantum local‑area networks (QLANs). Existing pure‑quantum medium‑access‑control (MAC) schemes based on Dicke states can resolve contention within a single QLAN without any classical round‑trip communication, but they do not scale to hierarchical, dynamic environments where the number of winners and the sizes of QLANs vary over time.

DH‑EAC solves this by employing a two‑stage quantum lottery. In the outer stage, a global orchestrator (GO) prepares a Dicke state |Dₘᴷ⟩, where m is the number of QLANs and K is the number of QLANs that will be granted access for a given request. Each QLAN receives one qubit of this state; a local Z‑basis measurement yields exactly K “1” outcomes, uniformly selecting K QLANs from the m possibilities. Once the winning set S is known, a deterministic decision function g maps S (together with the per‑QLAN availability numbers nᵢ^{avail} and the total requested winners k_req) to a quota vector (k₁,…,kₘ). The function g is implemented coherently, either via a Quantum Read‑Only Memory (QROM) lookup for small instances or via reversible arithmetic for larger instances, ensuring that the quota assignment is fixed at the moment the outer measurement collapses.

In the inner stage, each winning QLAN i ∈ S locally prepares a Dicke state |D_{nᵢ}^{kᵢ}⟩, where nᵢ is the number of nodes currently able to participate in QLAN i and kᵢ is the quota assigned by g. Again, each node measures its qubit; exactly kᵢ nodes obtain outcome “1”, thereby becoming the winners within that QLAN. Because all decisions are resolved by quantum measurement alone, no classical messages are required to adjust the outcome, eliminating round‑trip latency that would otherwise cause decoherence or expose participant identities.

The authors formalize anonymity as the property that no node learns the identity of any other winner before the measurement is completed. Since GO only receives aggregate availability numbers and each LO only learns its own measurement outcome, the protocol preserves anonymity both across QLANs and within each QLAN. Fairness is quantified using Jain’s fairness index applied to the vector of per‑node winning probabilities. The deterministic quota rule g is designed to compensate for heterogeneous QLAN sizes, preventing larger QLANs from monopolizing the entanglement budget.

A performance model is built on an i.i.d. loss assumption, distinguishing loss probabilities (q_WAN, q_LAN) and transmission‑time constants (t_WAN, t_LAN) for the wide‑area and local links respectively. The model yields closed‑form expressions for success probability, expected latency, and throughput as functions of m, K, the quota vector, and the loss parameters.

Simulation experiments compare DH‑EAC against two baselines: (1) a single‑QLAN Dicke‑state MAC (the original pure‑quantum scheme) and (2) a classical GO‑driven allocator that uses classical signaling for winner selection. Scenarios vary the number of QLANs, the distribution of nᵢ^{avail}, the requested total winners k_req, and the loss rates. Results show that DH‑EAC suppresses latency growth as the number of QLANs increases, maintains higher success probabilities under realistic loss conditions, and achieves Jain’s index values close to 1 even when QLAN sizes are highly heterogeneous. Throughput scales favorably compared with the classical baseline, especially when WAN loss dominates.

Implementation considerations focus on the preparation of multipartite Dicke states and the cost of coherent quota lookup. The paper cites recent low‑depth Dicke‑state generation circuits and QROM optimizations that keep T‑count and circuit depth manageable for moderate m and nᵢ. Nevertheless, scaling to very large networks will require further advances in error‑corrected state preparation and efficient reversible arithmetic for the quota function.

In conclusion, DH‑EAC extends pure‑quantum contention resolution to a hierarchical, dynamic setting, delivering anonymity, fairness, and low latency without any classical round‑trip communication. The work provides a concrete protocol specification, analytical performance models, and simulation evidence, establishing a viable design point for entanglement access control in future wide‑area quantum internet deployments. Future directions include experimental demonstration of the required Dicke‑state preparation on photonic or superconducting platforms, integration with higher‑layer routing and scheduling algorithms, and exploration of multi‑request concurrent scheduling.