Quantum networks, or in their global form Quantum Internet, are one of the key emerging quantum technologies along with quantum cryptography, computing, and sensing. A quantum network is expected to provide a platform where users can exchange quantum information. Such capability is essential for quantum cryptography, distributed quantum computing, and quantum sensor networks. Although these are important applications on their own and there are various proposals to realize them, the exact utility and implementation details of quantum networks are still an active area of research, and there are many opportunities for innovation at various layers of quantum networks. This constitutes part of our research efforts at Cisco Quantum Lab. What is clear at this stage is that quantum information is to be transmitted across quantum networks in the form of electromagnetic waves. Here, we consider digital quantum systems where the information is represented in terms of quantum bits or in short qubits. So, the core purpose of a quantum network is to allow its users to exchange qubits in the form of photons (or tiny wave packets of light).

A basic challenge in enabling future quantum internet is to overcome signal degradation during transmission over long distances. Unfortunately, fundamental laws of quantum mechanics forbid the applicability of classical solutions to quantum networks, and novel ideas based on quantum need to be developed. In a recent project, we started a new line of research in this direction which we will expound further in this post.

Over the past twenty years or so, numerous protocols which are generally referred to as quantum repeaters (named after their classical counterpart) have been devised to deal with the signal degradation across future quantum networks. The basic idea is to place a number of repeater stations at intermediate distances to effectively account for the photon loss.

Quantum repeater protocols must not put quantum-guaranteed security or privacy at risk by introducing new vulnerabilities. For instance, simple repetition schemes undermine quantum security. Consider a typical quantum communication where qubits are exchanged, and the security is guaranteed by the no-cloning theorem [which states that an arbitrary unknown quantum state cannot be copied unless the qubit is measured and the state is revealed]. As a result, the users will find out if an eavesdropper intercepts, since copying/observing arbitrarily unknown qubits requires measuring them. Now, consider the case where the sender sends multiple copies of the same qubit hoping that one will arrive at the destination. What is nominally regarded as lost qubits by the recipient, in this case, might have very well been taken away by an eavesdropper. Therefore, quantum repeaters require new technologies and quantum operations beyond classical repeaters. In what follows, we review some hardware issues and our recent efforts at Cisco Quantum Lab to find possible ways to address them.

Quantum repeater protocols are generally divided into two categories in terms of the type of required communications:

1. Two-way repeaters: These protocols are based on heralded quantum entanglement distribution, where a pairwise entanglement link (Bell pair) between the sender and receiver is established in a three-step procedure as shown below. First, short-range Bell pairs between adjacent repeaters are formed. Second, the qubits at intermediate nodes are measured in the local Bell basis (this process is also known as entanglement swapping). Finally, the measurement outcomes are announced to the neighboring repeaters, which requires a two-way communication channel. The quantum information is then teleported via the generated Bell pair.

Figure 1: Two-way repeater

2. One-way repeaters: These protocols are based on quantum error correction, where encoded quantum information is transmitted in the form of multi-photon states. These encoded states are resilient against photon loss up to a certain number which depends on the quantum code; roughly speaking, the more photons used to encode a qubit the more resilience we get. Intermediate repeater stations then check the incoming state for errors and prepare a fresh encoded qubit as the output to be sent to the next repeater. Therefore, the information is always transferred in one direction, and the encoded qubit is read off at the destination.

[caption id="attachment_5033" align="aligncenter" width="690"] Figure 2: One-way repeater[/caption]

The two-way communication required in the first category often leads to new challenges at scale such as latency and long-lived quantum memories (usually run at cryogenic temperatures) at each station and network congestion may occur. For these reasons, we mainly focus on one-way repeaters in our recent work, where we put forward a general framework that involves only photons, and hence named all-photonic one-way quantum repeaters. This means that in principle there is no need for quantum memory. Furthermore, compared to the previous literature we base our repeater protocol on three principles: simplicity, flexibility, and efficiency, as we explain further below.

[caption id="attachment_5035" align="aligncenter" width="613"] Figure 3: All-photonic one-way repeater[/caption]

• Simplicity- Past literature often involves performing some sort of quantum error-correcting operation at each repeater node (as shown in Figure 2) which can be quite complicated. In contrast, our protocol simplifies this process by only applying fixed quantum gates and direct detection, thereby postponing all data processing and error correction to the receiver. This will simplify the hardware and software at repeater stations.

• Efficiency- Previous protocols use many photons (hundreds to thousands) per encoded qubit. Generating such large codes is onerous since it is difficult to maintain a large number of photons coherent. The fact that we do error correction at the receiver’s location makes it possible to apply distributed algorithms to correct errors across the network and leads to a more efficient implementation, i.e., similar performance with significantly fewer photons.

• Flexibility- A common theme in conventional repeater protocols is that the hardware is entirely designed for specific quantum codes and substantial changes need to be made to switch to another code. In our work, we bring all the complexity into a single device, the so-called resource state generator (RSG), which outputs multi-photon encoded qubits by sending pulsed laser through an array of interferometers and photon detectors (see Figure 3 above). RSGs are fabricated on (silicon) photonic integrated circuits which are also a candidate platform for quantum computing (though we do not need the full quantum computational capability). This way, upgrading to a new quantum code would become a software update, i.e., reprogramming the photonic circuit. Such flexibility further leads to a long-term advantage as new generations of quantum codes will be available. For example, one can leverage the remarkable properties of the recently developed quantum low-density parity check (QLDPC) codes in our repeater scheme.

In conclusion, there are many challenges in developing quantum repeaters to enable wide-area quantum networks. This means that there are also many opportunities for innovation. At Cisco Quantum Lab, we believe that leveraging recent breakthroughs in integrated quantum photonics to build novel flexible architectures for quantum repeaters is a promising path forward.