Differentiated Service Entanglement Routing for Quantum Networks

TL;DR

This work tackles entanglement routing in scalable quantum networks under limited entanglement resources. It introduces the Differentiated Service Entanglement Routing (DSER) framework, combining a tensor-based lowest-loss path search with a differentiated First Request First QoS (FRFQoS) channel allocation to serve multiple users with distinct QoS requirements. The method maintains SP, SC, and a QoS queue to maximize active entanglement connections while prioritizing earlier requests, and it is evaluated on TDM/WDM network configurations showing higher key rates than prior approaches. The results demonstrate DSER’s potential for large-scale, configurable quantum networks and real-time dynamic request handling, with future work extending to quantum memory and repeater integration for richer QoS guarantees.

Abstract

The entanglement distribution networks with various topologies are mainly implemented by active wavelength multiplexing routing strategies. However, designing an entanglement routing scheme, which achieves the maximized network connections and the optimal overall network efficiency simultaneously, remains a huge challenge for quantum networks. In this article, we propose a differentiated service entanglement routing (DSER) scheme, which firstly finds out the lowest loss paths and supported wavelength channels with the tensor-based path searching algorithm, and then allocates the paired channels with the differentiated routing strategies. The evaluation results show that the proposed DSER scheme can be performed for constructing various large scale quantum networks.

Figures (14)

• Figure 1: The architecture of the configurable quantum entanglement distribution network. The network can be divided into the physical layer, the control layer and the application layer. The user's request for establishing or deleting entanglement connections will be transmitted to the control layer. The controller performs the routing strategies according to different user requirements and generates the entanglement flow tables to the physical layer. Afterwards, the physical layer configures the forwarding devices according to the entanglement flow tables. QKD is short for quantum key distribution, QDS is short for quantum digital signature and QCS is short for quantum clock synchronization.
• Figure 2: Schematic diagram of the DSER scheme. $\mathrm{Req}=[\nu_i,\nu_j,opt]$ means the routing request between user $\nu_i$ and $\nu_j$ and $opt\in [\mathrm{Establish}, \mathrm{Delete}]$. The network graph $G$, the lowest loss path set SP, the channel allocation set SC and the QoS priority queue Q are stored in the network status database. The blue arrow indicates the data transmission lines and the black arrow indicates the flow control lines.
• Figure 3: Schematic diagram of the proposed Tell-PS algorithm. $L=\{l_r^\beta\}, P=\{p_r^\beta\}, U=\{u_r^\beta\}, r=1,2,\cdots,k$ and $\beta=1,2,\cdots,m_r$, $k$ is the total count of nodes, $m_r$ is the total supported port count of the node $\nu_r$. $l_r^\beta$ is the link loss between $\nu_s$ and the $\beta$-th port of $\nu_r$, $p_r^\beta$ is a vector composed of the edge from $\nu_s$ to the $\beta$-th port of $\nu_r$. If the $\beta$-th port of $\nu_r$ is not directly connected to the node $\nu_s$, $p_r^\beta=\emptyset$ and $l_r^\beta=\infty$. $u_r^\beta$ represents the visit status of the $\beta$-th port of $\nu_r$. $(\nu_q,\alpha)=GPL(L,U)$ returns the $\alpha$-th port of next node $\nu_q$ with the lowest loss. $w_{\alpha,\alpha,:}^q$ represents the $\alpha$-th port input and $\alpha$-th port output loss vector of pass-through loss tensor $\mathcal{W}_q$ for node $\nu_q$. $C_{qr}[:, \beta]$ represents the $\beta$-th column of $C_{qr}$. $[m,m_{row}]=min(Z)$ represents getting the minimum value $m$ and index number $m_{row}$ in $Z$.
• Figure 4: Secure key rate vs brightness $\mathrm{B}$ for different user loss $l_i,l_j$, coincidence window $t_{CC}$, and polarization measurement errors $e^{pol}$, dark count rate $\mathrm{DCR} = 300\mathrm{cps}$, and the brightness of each wavelength channel is $3.75\times 10^7\mathrm{cps}$. The red solid label indicates the optimal count of channels that achieve the highest secure key rate.
• Figure 5: First request first QoS routing strategy. $ts_i$ means the request time series and $ts_{i+1}>ts_i$. $\mathrm{Req(n)}=[\nu_i,\nu_j,opt]$ means the user request, $n=1,2,3,4$. $t_{\mathrm{CC}}$ is the coincidence window, $e^{pol}$ is the individual polarization error probability and dark count rate $\mathrm{DCR}=300\mathrm{cps}$. The solid red label indicates the optimal channel count in (A)-(D). (E) indicates the channel allocated result.

Theorems & Definitions (12)

• Definition 3.1
• Definition 3.2
• Definition 3.3
• Definition 3.4
• Definition 3.5
• Definition 3.6
• Definition A1
• Definition A2
• Definition A3
• Definition A4