Quantum Network Architecture and Protocol Stacks 

In QUANT-NET, we consider a quantum network that can support multiple concurrent users and multiple quantum nodes. Essentially, the quantum network consists of four major types of quantum entities as shown in Figure 5 below: 

• Quantum end nodes (Q-nodes) 

• Quantum repeater (QRs). 

• Bell-State-Measurement nodes (BSM-nodes) 

• Quantum channels. 

Q-nodes, QRs, and BSM-nodes are connected to optical switches through optical fibers. The optical switches are further connected among one another to form a meshed all-optical network. Dedicated wavelengths of these fibers are used as quantum channels to transmit quantum signals between Q-nodes, QRs, and BSM-nodes. Through dynamic provisioning, multiple logic quantum networks can be generated from the same underlying physical network. 

Figure 5: QUANT-NET Network Model

QUANT-NET uses a logically centralized control style and decouples the control and data plane. In the control plane, one or multiple quantum network controllers monitor the status of the quantum network. Quantum network server(s) run on top of the controller to perform control and management functions. In the data plane, quantum signals and messages are transmitted across.

Figure 6: Quantum networking architecture 

Inspired by the TCP/IP architecture, QUANT-NET implements a similar layered quantum networking architecture (see Figure 6), which describes how quantum network functions are vertically composed to provide increasingly complex capabilities. The layered quantum networking architecture relies on five key vertical layers:

• Quantum physical layer. This layer deals with the physical connectivity of two communicating quantum nodes.
  
  

• Quantum link layer. The quantum link layer handles robust entanglement generation between neighboring nodes (e.g., Q-node, BSM-node, and QR). Key error mitigation mechanisms, such as heralded entanglement generation (HEG), and heralded entanglement purification (HEP), have been proposed to suppress loss and operation errors in quantum networks. QUANT-NET makes use of these advanced mechanisms to build quantum link layers. The quantum link layer functions are implemented distributively between Q-nodes, BSM-nodes, and QRs at the control of the quantum network server.
  
  

• Quantum network layer. This layer is responsible for entanglement routing, which selects a path in the network along which to establish end-to-end entanglement links between a Q-node pair by performing entanglement swapping at intermediate QRs. It also selects intermediate nodes along the path to perform entanglement purification if necessary. The quantum network layer functions are mostly implemented in the quantum network server. After the path has been successfully chosen, the QRs along the path perform low-level functions, such as entanglement swapping, entanglement purification, or quantum error correction, at the control of the server.
  
  

• Quantum transport layer. This layer provides an end-to-end quantum signal transfer service to users and applications. It handles end-to-end entanglement generation and entanglement purification to provide high-fidelity entangled pairs between Q-nodes. Quantum signals are transferred by teleportation. Quantum QoS requirements such as fidelity and entanglement generation rate are specified by the high-level quantum application layer. The quantum transport layer functions are implemented in Q-nodes at the control of the quantum network server. 
  
  

• Quantum application layer. The application layer provides the interfaces and protocols needed by users of a quantum network. Through the interfaces, quantum transport endpoints can be specified, and quantum QoS requirements such as fidelity and entanglement generation rate can be negotiated or specified.

The above layered quantum network architecture and protocol stack function assignments are tentative and experimental. When research has been conducted, lessons will be learned and experiences will be gained. We may then determine whether new layer(s) and/or function(s) need to be added, or whether existing layer(s) and/or function(s) need to be changed or modified.