The first device that relies on quantum mechanics to create unhackable computer codes was unveiled earlier this month. The Navajo Secure Gateway launched on 3 November in the US, and similar products yet to be launched, prove it is possible to use the most intriguing of physical laws for encrypting information. Now the race to be first to market is over, a new question arises over how practical this first generation of devices will actually be.
The big snag with quantum cryptography is the relatively limited distance over which such encoded information can be successfully transmitted. This problem is the focus of work being carried out by Toshiba’s Quantum Information Group, led by Dr. Andrew Shields at the company’s Cambridge laboratories.
Shields is working with quantum information, applications using quantum mechanics in information technology – an area on the far reaches of scientific understanding.’Quantum mechanics is the theory of the physical world describing theuniverse and the physical things in it,’ explained Shields.
‘It has very successfully described things such as atoms, molecules and how things such as semiconductors work. A couple of years ago people realised it could also be applied to information. That was quite a strange idea, but a lot of exciting things have come out of that, quantum cryptography being one of them.’
Encryption keys used to scramble and secure data are vulnerable to being copied or broken, but a quantum-based system will offer companies guaranteed protection using uncrackable codes. Such technology is highly sought after by organisations such as government agencies and banks, which regularly transfer valuable or sensitive data between sites.
‘With key cryptography you keep your key for an extended period of time,’ said Shields. ‘Hacking can do a lot of damage and you have no way of telling that it is being done. Public key infrastructure (PKI) security requires complex key management using a database of codes, meaning you have to place your trust in the database operator. With quantum cryptography, you can authenticate and identify users as well as protect your data. Code-based security methods are based on assumptions that certain mathematical operations are difficult, but there is no promise that they are impossible.’
However, quantum cryptography is currently limited to sending inform-ation over relatively short distances of around 100km, something that Shields is working to resolve. That inform-ation is transmitted as streams of photons.
As the length of the fibre used for data transfer gets longer, the more photons are scattered from it, meaning the number that survive long journeys can be hard to distinguish from the background noise. Over distances of under 100km, the natural error rate this causes is between three and seven per cent, but as the distance increases this shoots up fairly rapidly. Noise then starts to dominate over the actual signal, increasing to a level where it causes an error rate of over 11.5 per cent, the point where natural errors become indistinguishable from those caused by an eavesdropper.
To counter this, Shields and his team have developed an ultra-low noise detector, allowing messages to be sent over much longer distances. And this year the Toshiba team broke all records by becoming the first to send quantum-encoded data over 100km. It anticipates that using existing equipment the maximum distance achievable will be between 130 and 140 km.
Recognising this limit, the team is now working to further improve the technology by developing a light source that releases only a single photon. At present, photons are generated using a pulse laser diode. However, each pulse may only carry a photon in one tenth of the cycle, which is very inefficient and has a detrimental effect on how much data the fibre can carry and for how far. It also has security implications.
‘Sometimes in each cycle two photons may be generated, and this can put data at risk of a photon-splitting attack,’ said Shields. ‘Hackers could take one photon, leaving the other to continue down the fibre, compromising the key. This type of attack is very hard to carry out, but in future it could become a reality.’
The single photon source uses quantum dots consisting of indium arsenide grown on a layer of gallium arsenide, covered by an opaque film with an aperture. This gap allows only one photon to escape each time the device is excited and can be made in the same way as a normal LED, making manufacture affordable. It is also compatible for use with normal optical fibres, and should be ready for release within the next two to three years.
Shields jokingly refers to the resulting single photon-emitting diode as ‘the world’s dimmest LED’.
‘It doesn’t sound useful, but for quantum cryptography and the like, it really is,’ he said.
As with all cutting edge work Toshiba faces competition – in this case from New York-based MagiQ Technologies and Switzerland’s id Quantique, a spin-off from the University of Geneva. Both have announced that they can produce working systems ready for market. However, as Shields explained, the team is avoiding involvement in this particular race, with the hope of trumping their rivals by unveiling a better quality system in the future.
‘Both the MagiQ and id Quantique systems work in a different way to ours,’ said Shields. ‘We are putting our efforts into the development of device technology as this determines the eventual performance of the system. Its success will depend upon this. Luckily we do not have the same pressure to impress investors by pushing out a working system before the market is ready to buy it. Personally I think it is a bit premature to be selling systems. It takes time to introduce customers to such things and it will probably still be about three to four years before quantum cryptography is widely accepted, despite the fact that it is already reliable enough to use.’
Toshiba’s latest achievement has been to make the system easier to install within the average IT room, integrating the technology into a video-sized box that can be racked with other equipment. A prototype should be ready for release within the next few months.
Although there is still much work to do, the team is already turning to its next challenge, quantum computing.
‘Quantum cryptography will be the first application of quantum mechanics that will come to the market, but we are gradually dealing with quantum computing, too,’ said Shields. ‘However, this is much more technologically challenging.’
Though he recognises that it will be a long time before any form of workingsystem will emerge, Shields is confident that the group’s work will prove vital.
‘It is 50 years since the first transistor was developed by Bell Laboratories and we have come a long way since then. I have the feeling that one day people may look back at our quantum work and see it as being in the same basic state as those first attempts, but also as an important step towards future systems.’
Living by a new code
Digital information is usually coded via the public key encryption system. This relies on mathematical codes with a 56-256 bit value to scramble data that is then decoded using an algorithm key. The sender encrypts the message with a public key, and the receiver decrypts it using a private key. Though mathematical keys are becoming more complex, code breakers have shown they can be copied or broken, given time and a powerful computer.
Quantum cryptography allows secure communications on fibreoptic networks. Uniquely it offers ‘unconditional secrecy’ which is independent of the computing power, fancy gadgetry or guile of an adversary. The security of quantum cryptography stems from inviolable laws of nature and it is regarded as the strongest cryptographic protection possible.
Many of the basic rules of quantum mechanics appear to be counter to our classical understanding of the world. An example of this, which is the basis of quantum cryptography, is that in general measurement alters the state of an unknown single quantum. This fact is the essential reason why a third party cannot copy a key sent by quantum cryptography.
It can be implemented by sending an encoded single photon along a standard telecom line. The sender encodes one bit, i.e. 0 or 1, on to each photon, which can be done in a number of ways. Because the information is carried by a single photon it is not possible for a hacker to tap in and remove part of the signal. Single photons do not split, so if the hacker measures the photons carried by a fibre, they will not reach the intended recipient. Only the photons that arrive at their intended destination are used to form the code key.
To conceal their presence, hackers might attempt to measure the photons and retransmit them. This will also fail. Quantum mechanics says that encoded photons cannot be copied faithfully. Trying to read and copy them will introduce errors into their encoded values, again alerting the recipient to the interception.