Quantum Communication Technology

Nicolas Gisin and Rob Thew

Group of Applied Physics, University of Geneva, 1211 Geneva 4, Switzerland
(Dated: October 25, 2018)
Quantum communication is built on a set of disruptive concepts and technologies. It is driven
by fascinating physics and by promising applications. It requires a new mix of competencies, from
telecom engineering to theoretical physics, from theoretical computer science to mechanical and
electronic engineering. First applications have already found their way to niche markets and uni-
versity labs are working on futuristic quantum networks, but most of the surprises are still ahead of
us. Quantum communication, and more generally quantum information science and technologies,
are here to stay and will have a profound impact on the XXI century.
arXiv:1007.4128v1 [quant-ph] 23 Jul 2010

Introduction: Quantum Communication enjoys an en- tion science and technology - is that it exploits quantum
viable position in physics, in between fundamental quan- physics at the level of individual quanta.
tum mechanics and applied quantum optics [1]. For most The simplest example is the quantum random number
physicists, quantum communication is merely a play- generator. Since the detection of a single photon after
ground to explore fascinating topics like entanglement, one of the two output ports of a beam splitter is an in-
superposition of large objects, and, more generally, to trinsically random event, it offers a valuable source of
look for places where quantum physics may fail, that randomness, see Figure 1. Moreover, according to to-
is to explore the limits of quantum physics. This play- day’s physics, such a source of randomness is unique: no
ground requires new technologies and concepts. Usually, device based on classical physics will ever produce true
new technologies are driven by applications and quantum randomness, only at best ”pretty good” pseudo-random
communication is no exception: the emerging and future numbers, or noise, whose origin is hard to fully identify.
technologies are driven by the need for Yet, engineering a photon source, beam-splitter and two
CMOS-based single-photon detectors is not that com-
1. Fast Quantum Random Number Generators plex: the existing commercial QRNG is about the size of
(QRNG): from cryptography to internet lotteries a match-box.
and gaming,

2. Reliable fiber-based Quantum Key Distribution “0” “1”

(QKD): for today’s cryptography applications,

3. Quantum repeaters: for future continental scale

1 1 1 0 0 1 Key Generation
fiber optic quantum communication,

4. Earth to satellite links: for free space quantum FIG. 1: Quantum random number generation - one single
communication. photon at a time is sent to a 50/50 beamsplitter and can only
exit in one of the output modes. This process is fundamentally
The first two are already commercially available [2] and random and the photon’s detection is used to the generate
have already found small niche markets, while the later truly random bit strings.
two are still at an early stage.
A Conceptual Revolution: The basic idea of quantum
communication is to take advantage of the oddities of Another example, whose basic principle is rather
quantum physics, like the uncertainty relation, the su- straightforward, is QKD [3]. Every first year quantum
perposition principle and randomness. Note the concep- physics student knows that measurements tend to un-
tual revolution: instead of being afraid of quantum pecu- avoidably disturb the quantum state of the system under
liarities and trying to avoid their detrimental effects for investigation. This has puzzled generations of students
standard technologies, the new generation of quantum and professors. Now, if this ”negative fact” is applied
engineers aim at exploiting the new physics. In partic- to an adversary, like a spy on a communication chan-
ular they fully admit quantum physics as it stands and nel where the bits are carried by quantum systems, like
want to find original uses for its most counter-intuitive photons, then it is the spy who can’t measure the bits
features. It is somewhat surprising, and disappointing, without unavoidably leaving a trace of her intrusion un-
that it took six or seven decades before realizing that der the form of some disturbance. Again, the reasoning is
this new physics ought to produce new technology. One so simple that one wonders why no student came across
might argue that the laser, semiconductors, superconduc- that idea long ago (or did their professor tell them to
tivity, among others, are technologies based on quantum shut-up and compute?). Thus, while QRNGs generate
physics. However, the big difference with quantum com- randomness, QKD provides a means of distributing pri-
munication - and more generally with quantum informa- vate (secure) randomness.
 2

Further presentation of what quantum communication must. The best detectors in terms of efficiency are su-
is and how it relates to entanglement and other quantum perconductor bolometers, though these are prohibitively
oddities can be found, e.g. in [1]. In the following we slow and operate at a few milliKelvin [15]. QKD appli-
concentrate on future technology challenges. cations need cheap, compact, electronically cooled detec-
Near-Future Technologies: The most advanced appli- tors. Today this is achieved with semiconductor detectors
cations of quantum communication are clearly QRNG (InGaAs APDs) though their performance and function-
and QKD. The first was initially developed as a compo- ality needs to greatly improve. The APDs need to have
nent of the second. It was originally thought that QRNG lower dark count rates and ”afterpulsing” [16], which can
would also find applications in classical cryptography and introduce errors on the key. Furthermore, one of the most
in Monte-Carlo numerical simulations. QRNG did find important characteristics of single-photon detectors, that
some application in classical cryptography (e.g. the state is too often neglected, is the maximum count rate; future
of Geneva uses QRNG to produce the pin-codes used for QKD systems will need to generate several Mbps of se-
internet voting), but by far the largest application came cure keys.
as a surprise: internet gambling for which it has now been Some physicists speculate that QKD systems using
certified by Metas [4] This is a good example that appli- not-so-weak laser pulses and homodyne detection, con-
cations of new technologies are hard to predict (physicists tinuous variable (CV) QKD [17], will outperform single-
are especially bad at predicting good commercial appli- photon schemes. They argue that, contrary to single-
cations!). photon schemes, homodyne detection always produces a
QRNG development is a timely topic, but any ap- result. This is correct, though the results are necessarily
proach should concentrate on three key requirements: very noisy. We expect that it is more efficient to let Na-
ture select the cases with low noise, i.e. the cases where
1. Origin of randomness easily identifiable. One a single-photon is detected, rather than to always have a
should be able to quantify how much randomness noisy result, where the noise has then to be removed by
is truly quantum and how much is ”technological sophisticated error correction algorithms. Furthermore,
noise”, e.g. thermal noise, detector noise, etc. for long distances the not-so-weak laser pulses tend to
become pseudo-single-photon and the difference between
2. Reliability, size and price. There is no fundamental the two systems vanishes. But, admittedly, the future
reason for a QRNG not to be as small and cheap will show us the truth with possibly both systems find-
as a standard electronic chip. ing their niches.
An increasingly important requirement for future QKD
3. Fast, in particular faster than classical, physical,
schemes is that they run on the same fiber as the classi-
RNG. The minimal rate of future QRNG should
cal channels (both the classical processing and encrypted
be in the range of hundreds of Mbps to Gbps.
data channels). This is a serious challenge as the inten-
Present QKD systems are mostly based on the his- sity difference is huge: 8 to 9 orders of magnitude. Hence,
torical BB84 protocol [5] (with some improvements like Raman scattering and other nonlinear effects have to be
SARG [6] and Decoy-state [7–10]). However, better pro- taken into account: even microwatts can produce enough
tocols have been invented in the context of fiber networks photons to impair the quantum communication; recent
[11, 12] . It should be understood that BB84 originally efforts suggest that multiplexing quantum and classical
was described using polarization encoding, which is intu- channels in a fiber is limited to around 50 km [18] with
itively easy to understand, though in practice most real current technology.
systems use some type of phase encoding that is more A serious push towards network compatibility can also
compatible with fiber optical systems. Despite these ad- be witnessed by the number of QKD test beds running or
vances, the best QKD protocols have probably not yet planned worldwide. In 2008 in Vienna the European con-
been discovered. In any case, the protocol should use sortium SECOQC demonstrated a mix of different QKD
telecom photons (i.e. around 1550 nm), be compatible systems running in a complex network [19]. A triangular
with standard optical fiber networks and combine them network has been running continuously in Geneva (data
with the necessary quantum features to guarantee ”quan- available in real time at www.swissquantum.com) since
tum security”. This requires synergy between telecom April 2009. In Durban, South-Africa, yet another net-
engineers and quantum theorists. work runs continuously carrying real data and in October
Most technology developments on QKD concentrate 2010 a large network will commence operation in Tokyo,
on single-photon sources and on detectors [13]. However, while others have been announced for Madrid and China.
somewhat surprisingly, single-photons are not required Another sign of the maturing of QKD is the appear-
for QKD: it is much easier to use so-called pseudo single- ance of quantum hackers. They do not attack the prin-
photon sources, i.e. strongly attenuated pulsed lasers. ciple on which QKD relies, as this is provable secure,
These are cheap, very reliable and fast (GHz rates). Note but take advantage of implementation weaknesses [20–
however, that single-photon sources could find their ap- 22]. The latter can, and have to be, tested and strength-
plication in quantum repeaters [14], see below. Improv- ened, rendering QKD more and more reliable.
ing single-photon detectors, on the contrary, is a real The rate of future QKD systems will be such that true
 3

Mbps one-time pad encryption should be possible over huge. First, contrary to point-to-point QKD, quantum
metropolitan networks. This is including all the real-time repeaters require entanglement, i.e. the ingredient neces-
classical processing, communication and network over- sary for quantum teleportation. Hence, mere attenuated
heads. It will thus be the result of an interdisciplinary laser pulses will never suffice. Next, it is crucial that
team of engineers. one distributes entanglement between neighbor stations
Future Technologies and Applications: This approach in parallel and stores it until two neighboring stations
to fiber based QKD is ultimately limited in distance to are entangled. This required quantum memories [25], i.e.
a few hundred kms. Indeed, for 1000 km, even with a the capacity to store photons, or more precisely their
perfect 10 GHz single-photon source, ideal detectors and quantum state, in a reversible way without loosing any
0.2 dB/km fiber losses one would detect only 0.3 photon of their quantum features; in particular quantum memo-
on average per century! Consequently, futuristic conti- ries ought to preserve entanglement. Furthermore, it has
nental scale quantum communication requires completely recently be shown that ”reasonable” rates (≥ 1bps over
different technologies than today’s QKD systems, mere 1000 km) are possible only if the quantum memories are
improvements will not do. There are two main paths: vastly multimode, i.e. able to store hundreds of quantum
satellite-based and quantum repeaters. bits simultaneously.
Satellite-based quantum communication is conceptu- Each of the necessary ingredient for a quantum re-
ally similar to fiber-based QKD, except that instead of peater have been individually demonstrated in various
fibers one sends the photons through free space between labs. However, not always with compatible technology
satellites and earth-based stations, both equipped with and not with sufficiently high specifications. However,
telescopes. Since no fibers are used, the choice of wave- there are good reasons to be optimistic. Firstly, as the
length is compatible with silicon APDs. The technologi- challenges are truly fascinating, some of the best stu-
cal requirement are not so much set by quantum optics, dents are doing their PhDs on quantum repeaters. Fur-
but are defined by the stringent specifications imposed thermore, Europe and several national funding agencies
by space agencies. Currently, ESA is investigating this have realized the potential and are thus supporting the
possibility [23] and the US and China also have programs research [26, 27].
in place. We wouldn’t be surprised if China is the first to Conclusion: Quantum communication technologies in-
demonstrate a QKD link between earth and a satellite. volve diverse disciplines, ranging from pure engineering
The alternative to satellites for continental distance problems (integrate a QRNG, optimize QKD systems),
quantum communication exploits a beautiful idea: quan- to fascinating basic physics (quantum memories and tele-
tum repeaters. Repeaters for classical optical communi- portation), via theoretical physics (design more efficient
cation, e.g. based on Erbium fiber amplifiers, are well and secure QKD protocols) and computer science (design
known, but unusable for quantum communication as any and implement the necessary processing for the raw data
stimulated emission process necessarily brings with it delivered by QKD) and lots of electronic, telecom and
spontaneous emission. At the single-photon level this software engineering (the real work). The market is still
noise is as strong as the signal, hence standard repeaters small, but burgeoning, the physics fascinating, the chal-
don’t work in the quantum regime (this is a form of the lenges mind-boggling. Only one thing is certain: there
quantum no-cloning theorem that follows from the linear- will be surprises.
ity of Schrödinger’s equation). The basic idea of quan-
tum repeaters is to first establish entanglement between
a series of stations, and next, to use quantum physics Acknowledgment
[24] to teleport a ”photon” (more precisely, the quantum
state carried by a photon) from one station to the next. This work has been supported by the EU projects QuReP,
As fascinating as this is, the technological challenge is Q-ESSENCE and QUIE2 T and by the Swiss NCCR-QP.

[1] Gisin N. and Thew, R., ”Quantum Communication”, Na- tum cryptography protocols robust against photon num-
ture Photonics 1, 165-171 (2007). ber splitting attacks for weak laser pulses implementa-
[2] www.idQuantique.com. tions”, Phys. Rev. Lett., 92, 057901 (2004).
[3] Gisin, N., Ribordy, G., Tittel, W. and Zbinden, H., [7] Hwang, W.-Y., ”Quantum key distribution with high
”Quantum cryptography”, Rev. Mod. Phys., 145 (2002). loss: toward global secure communication”, Phys. Rev.
[4] The Swiss national metrology institute, Lett., 91, 057901 (2003).
http://www.metas.ch. [8] Wang, X.-B., ”Beating the photon-number-splitting at-
[5] Bennett, Ch. H. and Brassard, G., ”Quantum cryptogra- tack in practical quantum cryptography”, Phys. Rev.
phy: public key distribution and coin tossing”, Int. conf. Lett., 94, 230503 (2005).
Computers, Systems & Signal Processing, Bangalore, In- [9] Lo, H.-K., Ma, X. and Chen, K., ”Decoy state quantum
dia, 10-12, 175-179 (1984). key distribution” Phys. Rev. Lett., 94, 230504 (2005).
[6] Scarani, V., Acin, A., Ribordy, G. and Gisin, N., ”Quan- [10] Harrington, J. W., Ettinger, J. M., Hugues, R. J.
 4

and Nordholt, J.R., ”Enhancing practical security of [19] Peev, M., et al., ”The SECOQC quantum key distribu-
quantum key distribution with a few decoy states” tion network in Vienna ”, New J. Phys, 11 075001 (2009).
quant-ph/0503002, Los Alamos report LA-UR-05-1156 [20] Gisin, N., Fasel, S., Kraus, B., Zbinden, H. and
(2005). Ribordy, G., ”Trojan-horse attacks on quantum-key-
[11] Inoue, K., Waks, E. and Yamamoto, Y., ”Differential- distribution systems”, Phys. Rev. A, 73, 022320 (2006).
phase-shift quantum key distribution using coherent [21] Makarov, V., Anisimov, A. and Skaar, J., ”Effects of de-
light”, Phys. Rev. A, 68, 022317 (2003). tector efficiency mismatch on security of quantum cryp-
[12] Stucki, D., Brunner, N., Gisin, N., Scarani, V. and tosystems”, Phys. Rev. A, 74, 022313 (2006).
Zbinden, H., ”Fast and simple one-way Quantum Key [22] Qi, B., Fung, C.-H. F., Lo, H.-K. and Ma, X., ”Time-shift
Distribution”, App. Phys. Lett., 87, 194108 (2005). attack in practical quantum cryptosystems”, Quant. Info.
[13] Hadfield, R. H., ”Single-photon detectors for optical Compu. 7, 43 (2007)
quantum information applications”, Nature Phot. 3, 696 [23] Aspelmeyer, M., Jennewein, T., Pfennigbauer, M.,
(2009). Leeb, W. and Zeilinger, A., ”Long distance quantum
[14] Sangouard, N., Simon, C., Minar, J., Zbinden, H., de communications with entangled photons using satel-
Riedmatten, H., and Gisin, N., Phys. Rev. A, 76, 050301 lites”, IEEE J. Sel. Top. Quantum Electronics, 9
(2007). 1541 (2005); see also the Space Quest programme at
[15] Lita, A. E., Miller, A. J., and Nam, S. W., ”Counting http://www.quantum.at/quest
near-infrared single-photons with 95% efficiency” Opt. [24] Bennett, Ch. H., Brassard, G., Crépeau, C., Jozsa, R.,
Exp., 16, 3032 (2008). Peres A., and Wootters, W. K., ”Teleporting an unknown
[16] An afterpulse is caused by trapped charges in the APD quantum state via dual classical and Einstein-Podolsky-
being released when the detector is reset causing another Rosen channels”, Phys. Rev. Lett., 70, 1895 (1993).
avalanche resulting in a false detection event. [25] Simon, C., et al., ”Quantum memories: A review based
[17] Grosshan, F. and Grangier, Ph., ”Continuous Variable on the European integrated project Qubit Applications
Quantum Cryptography Using Coherent States”, Phys. (QAP)”, Eur. Phys. J. D, 58 1 (2010)
Rev. Lett., 88, 057902 (2002). [26] See the QuReP website at http://quantumrepeaters.eu
[18] Eraerds, P., Walenta, N., Legre, M., Gisin, N., Zbinden, and those listed at http://qurope.eu/projects
H., ”Quantum key distribution and 1 Gbit/s data en- [27] For example in Switzerland, the NCCR - quantum pho-
cryption over a single fibre”, J. Lightwave Tech, 28, 952 tonics programme: http://nccr-qp.epfl.ch
(2010)