A cost-effective architecture for quantum cryptography has been demonstrated in which a single receiver positioned at a network-hub node is shared by many end users to exchange secret encryption keys. See Letter p.69
Keeping a secret has never been easy. Throughout history, ideas for encrypting messages have spurred people to come up with ways of breaking the keys that encrypt the messages. Perhaps the most famous example is the race that occurred during the Second World War between the German Enigma cipher machines and the British Colossus — the world's first electronic, digital computer. In the 1980s, the game changed with the invention of a cryptographic technique known as quantum key distribution (QKD)1 that uses the laws of quantum physics to guarantee secure communication. So far, however, QKD has been demonstrated only for point-to-point communications and relatively simple networks. On page 69 of this issue, Fröhlich et al.2 describe a method that brings the advantages of QKD to as many as 64 end users, who can share a quantum key, and thus a secret. The results are illustrative of the worldwide research progress towards a practical 'quantum Internet'.
Through the seminal research of Auguste Kerckhoffs in the nineteenth century3 and Claude Shannon in the twentieth century4, it is possible to reduce the problem of secure communication to the secure transfer of a secret encryption key. Unfortunately, neither Kerckhoffs nor Shannon explained how secure key distribution could be performed. Today, it relies on the unproven difficulty of cracking certain hard mathematical problems, such as factoring large integers. But factoring methods are continually improving, making the security lifetime of this method hard to predict.
Quantum cryptography avoids these issues. In this technique, the key is encoded into quantum states, such as the polarization of a series of single photons that are passed between two parties trying to share secret information. Heisenberg's uncertainty principle dictates that a third party trying to decode the key cannot look at these photons without changing or destroying the information they carry. In this case, it does not matter what technology the third party has: it will never be able to break the laws of quantum physics and decrypt the key.
More-recent research5 shows that QKD is secure, even if quantum mechanics turns out to be only an approximate theory describing the world. QKD is the first quantum-information application to reach the level of a commercial technology6,7. Present-day commercial QKD systems have been developed with a view to incorporating them into existing telecommunication infrastructures at the metro-area scale.
But we live in a networked world and QKD is intrinsically a point-to-point protocol. Several research groups have previously investigated how the advantages of QKD could be brought to multi-party networks (for example, see refs 8 and 9). These 'trusted QKD networks' amount to a mesh of point-to-point QKD links between nodes within which QKD-generated keys must be physically secured against adversaries (hence the need for the nodes to be trusted). However, this approach involves tremendous duplication of resources, with each node requiring QKD receivers to accept incoming photons, and QKD transmitters to send keys on to other nodes. The resulting high cost and limited scalability have been major obstacles to the adoption of QKD as a cyber-security technology.
In their study, Fröhlich et. al. introduce and demonstrate a cost-effective way to bring the advantages of QKD to multiple end users. This quantum access network, as they dub it, locates a single QKD receiver — with its expensive single-photon detectors — in a network hub, which serves multiple end users, each of whom has a QKD transmitter (Fig. 1). Each user's QKD photons are routed to the one receiver within the network hub over optical fibres. With the expensive single-photon detectors shared among many end users, the approach is cost-effective and reduces the hardware requirements of each user.
Fröhlich et al. show that their approach is scalable to as many as 64 end users, provided that the optical fibre between each user and the hub carries only the QKD single-photon signals. But in real-world access networks, these fibres are already 'lit up' in both directions with data-bearing optical signals, which introduce large numbers of photons that could degrade the QKD performance. However, a previous study10 using a single end user QKD transmitter demonstrated that QKD can successfully co-exist with this bi-directional optical traffic on an access network's optical-fibre links. Taken together with Fröhlich and colleagues' result, this provides strong evidence that QKD could be cost-effectively deployed as an overlay to existing fibre-to-the-home and other access networks.
The present study represents a perfect example of how conceptually new ideas can lead to new quantum technologies, and brings the advantages of a global quantum Internet a step closer to the consumer.
About this article
Optics Express (2014)