DODAG-X Protocol: Quantum Entanglement Routing

• DODAG-X protocol is a multipartite entanglement distribution scheme that uses a precomputed breadth-first spanning tree to efficiently generate GHZ states.
• It leverages a minimal eccentricity node to form an optimized DODAG structure, reducing classical and quantum overhead in path scheduling.
• The protocol achieves up to 35% fewer measurements, enhancing entanglement fidelity under realistic noise, loss, and network dynamics.

The DODAG-X protocol is a multipartite entanglement distribution scheme for quantum networks that combines the Destination Oriented Directed Acyclic Graph (DODAG) structure from classical networking with a variant of the X-protocol for graph-state entanglement routing. By precomputing a breadth-first spanning tree rooted at a node of minimal eccentricity, DODAG-X enables efficient, robust, and scalable entanglement generation under realistic noise, loss, and network dynamics. The protocol significantly reduces the quantum and classical overhead for generating Greenberger–Horne–Zeilinger (GHZ) states and can be extended to generic $n$-party entangled states (Negrin et al., 2024).

1. Network Model and Assumptions

DODAG-X operates on a quantum network whose physical topology is represented as an undirected graph $G_{\rm phys}=(V,E_{\rm phys})$, with nodes $v\in V$ denoting qubit-nodes (quantum processors or repeaters), and edges $e=(u,v)\in E_{\rm phys}$ as quantum channels (optical fiber, free-space link). The protocol establishes a virtual topology as a graph state $G_{\rm virt}=(V,E_{\rm virt})$, where each edge corresponds to an EPR-pair link created by controlled-Z ($CZ$) operations between qubits: $$\left|G_{\rm virt}\right\rangle = \bigotimes_{(u,v)\in E_{\rm virt}} CZ_{u,v}\ \left|+\right\rangle^{\otimes |V|}\,.$$ The protocol accounts for time-varying physical link failures, with nodes having up-to-date knowledge of their incident edges but not the global state, to avoid decoherence-inducing delays. Noise and fidelity models specify:

• Baseline gate and memory fidelity $F_0$ for each CZ and qubit,
• Decoherence acting independently with characteristic time $T_2$,
• Per-attempt link-loss probability $p_{\rm loss}$ and depolarizing qubit errors.

The fidelity of the generated entanglement after $G_{\rm phys}=(V,E_{\rm phys})$0 two-qubit operations and storage time $G_{\rm phys}=(V,E_{\rm phys})$1 is bounded by

$G_{\rm phys}=(V,E_{\rm phys})$2

highlighting the need to minimize both the number of measurements and the storage time.

2. DODAG Construction and Structural Properties

DODAG-X begins with root selection, choosing a vertex $G_{\rm phys}=(V,E_{\rm phys})$3 of minimum eccentricity

$G_{\rm phys}=(V,E_{\rm phys})$4

minimizing worst-case communication depth. A breadth-first spanning tree $G_{\rm phys}=(V,E_{\rm phys})$5 is constructed with directed edges $G_{\rm phys}=(V,E_{\rm phys})$6 oriented toward the root. Level assignments $G_{\rm phys}=(V,E_{\rm phys})$7 correspond to shortest-path distance to $G_{\rm phys}=(V,E_{\rm phys})$8 in $G_{\rm phys}=(V,E_{\rm phys})$9. Construction requires only a global BFS, with all subsequent entanglement routing via tree walks—obviating further path-finding or verification.

The structural overhead depends on the physical topology:

• On a grid $v\in V$0 ($v\in V$1), tree depth is $v\in V$2.
• For Watts–Strogatz small-world networks ($v\in V$3 nodes, mean degree $v\in V$4, rewiring $v\in V$5), expected depth is $v\in V$6.

3. The DODAG-X Protocol: Mechanics and Implementation

DODAG-X prescribes a scheduling of Pauli-X (and Z) measurements to convert the virtual tree state into a multipartite GHZ state among a chosen party set. For three parties $v\in V$7, the protocol proceeds as:

1. For each $v\in V$8, compute the unique path $v\in V$9 in $e=(u,v)\in E_{\rm phys}$0; locate intersections $e=(u,v)\in E_{\rm phys}$1 as the first common node of $e=(u,v)\in E_{\rm phys}$2.
2. For each party $e=(u,v)\in E_{\rm phys}$3, determine set $e=(u,v)\in E_{\rm phys}$4, select nearest $e=(u,v)\in E_{\rm phys}$5; along path $e=(u,v)\in E_{\rm phys}$6 (excluding endpoints), apply $e=(u,v)\in E_{\rm phys}$7 with $e=(u,v)\in E_{\rm phys}$8 as pivot via local-complementation.
3. For each distinct intersection $e=(u,v)\in E_{\rm phys}$9, traverse $G_{\rm virt}=(V,E_{\rm virt})$0 and alternately apply $G_{\rm virt}=(V,E_{\rm virt})$1 with the predecessor as pivot.
4. If $G_{\rm virt}=(V,E_{\rm virt})$2, apply $G_{\rm virt}=(V,E_{\rm virt})$3 using any $G_{\rm virt}=(V,E_{\rm virt})$4 as pivot.
5. For each $G_{\rm virt}=(V,E_{\rm virt})$5, remove all non-party neighbors via $G_{\rm virt}=(V,E_{\rm virt})$6.

This sequence yields a $G_{\rm virt}=(V,E_{\rm virt})$7 state on $G_{\rm virt}=(V,E_{\rm virt})$8. Each $G_{\rm virt}=(V,E_{\rm virt})$9 refers to a Pauli-X measurement (see Eq.~2.12 (Negrin et al., 2024)) implementing a local-complementation. Entanglement swapping and purification can be interleaved at repeater nodes by replacing logical X-measurements with Bell swaps and purification rounds.

4. Resource Optimization, Noise, and Fidelity

For three-party entanglement, total measurements $CZ$0 are partitioned into “path” and “isolation” types: $CZ$1 matching the count of the X-protocol on trees (Theorem 3.1, 3.2, A.3 (Negrin et al., 2024)). Since each Pauli measurement or local Clifford introduces infidelity, minimizing $CZ$2 optimizes $CZ$3, with

$CZ$4

A plausible implication is that DODAG-X's minimized measurement schedule directly improves practical achievable fidelities.

5. Computational Complexity and Classical Overhead

Classic X-protocols require on-demand shortest-path searches (Dijkstra/BFS) for each entangling event, incurring $CZ$5 or $CZ$6 complexity, which is prohibitive for large, dynamic, or densely connected quantum internets. DODAG-X amortizes this cost by a single BFS to build $CZ$7, with all subsequent routing and scheduling operations requiring only tree-walks of depth proportional to the shortest path to the root.

• Grid topology: each party-to-root path is $CZ$8–$CZ$9.
• Small-world topology: each path is $$\left|G_{\rm virt}\right\rangle = \bigotimes_{(u,v)\in E_{\rm virt}} CZ_{u,v}\ \left|+\right\rangle^{\otimes |V|}\,.$$0.

Theorems 3.1 and 3.2 establish (a) connectivity to intersections after Step 2 and (b) isolation of the multipartite entangled parties in the final state. Theorem A.3 establishes equivalence in measurement counts on trees versus X-protocol.

6. Quantitative Benchmarks and Empirical Results

Performance evaluation on two classes of topologies yields:

Grid Lattices ($$\left|G_{\rm virt}\right\rangle = \bigotimes_{(u,v)\in E_{\rm virt}} CZ_{u,v}\ \left|+\right\rangle^{\otimes |V|}\,.$$1)

$$\left|G_{\rm virt}\right\rangle = \bigotimes_{(u,v)\in E_{\rm virt}} CZ_{u,v}\ \left|+\right\rangle^{\otimes |V|}\,.$$2

DODAG-X achieves up to $$\left|G_{\rm virt}\right\rangle = \bigotimes_{(u,v)\in E_{\rm virt}} CZ_{u,v}\ \left|+\right\rangle^{\otimes |V|}\,.$$3 measurement reduction compared to the X-protocol, with only the $$\left|G_{\rm virt}\right\rangle = \bigotimes_{(u,v)\in E_{\rm virt}} CZ_{u,v}\ \left|+\right\rangle^{\otimes |V|}\,.$$4 grid incurring a minor $$\left|G_{\rm virt}\right\rangle = \bigotimes_{(u,v)\in E_{\rm virt}} CZ_{u,v}\ \left|+\right\rangle^{\otimes |V|}\,.$$5 penalty.

Small-World Networks ($$\left|G_{\rm virt}\right\rangle = \bigotimes_{(u,v)\in E_{\rm virt}} CZ_{u,v}\ \left|+\right\rangle^{\otimes |V|}\,.$$6, $$\left|G_{\rm virt}\right\rangle = \bigotimes_{(u,v)\in E_{\rm virt}} CZ_{u,v}\ \left|+\right\rangle^{\otimes |V|}\,.$$7, $$\left|G_{\rm virt}\right\rangle = \bigotimes_{(u,v)\in E_{\rm virt}} CZ_{u,v}\ \left|+\right\rangle^{\otimes |V|}\,.$$8)

DODAG-X saves $$\left|G_{\rm virt}\right\rangle = \bigotimes_{(u,v)\in E_{\rm virt}} CZ_{u,v}\ \left|+\right\rangle^{\otimes |V|}\,.$$9 measurements on average, up to $F_0$0 in highly clustered/high-rewired regimes. Weakly clustered/rewired instances experience a $F_0$1 increase in measurement counts.

These reductions directly translate into higher achievable fidelities and lower quantum resource consumption.

7. Multiparty Entanglement Extension and Generality

For three parties, DODAG-X produces a $F_0$2. The resulting graph after X-measurements is a triangle (3-star with central vertex removed). For $F_0$3,

• If all parties are on distinct branches of $F_0$4, a single-shot protocol applies (disjoint party-root paths except at $F_0$5).
• Otherwise, a multi-layer DODAG-X approach is used: group parties in threes, apply DODAG-X layerwise, then fuse roots using linear-cluster fusion techniques (as described by Fan et al. 2024).

Each measurement conjugates the underlying stabilizer group, with the final configuration realizing the stabilizer of a $F_0$6 state.

8. Comparative and Contextual Analysis

Relative to prior graph-state routing methods, DODAG-X provides:

• Efficiency: Eliminates repeated $F_0$7 path searches, replacing them with one $F_0$8 BFS and $F_0$9 tree-walks per entanglement event.
• Scalability: Path lengths scale as $T_2$0 (grids) or $T_2$1 (small-world) versus worst-case $T_2$2.
• Reliability: DODAG is maintained by classical control messages akin to RPL, enabling virtual entanglement overlay persistence under link churn, without additional quantum communication to update routing status.
• Resource Savings: Up to $T_2$3 fewer measurements in realistic topologies, which proportionally lowers exposed quantum operations and raises achievable quantum state fidelities.
• Flexibility: Universality for GHZ$T_2$4 generation in arbitrary topologies; naturally extends to larger GHZ$T_2$5 and magic-state routing within the graph-state formalism.

The marriage of the DODAG and X-protocol paradigms in DODAG-X establishes an efficient and noise-robust foundation for multi-user quantum communication in NISQ and early quantum-internet environments (Negrin et al., 2024).