IEEE - Aerospace and Electronic Systems - March 2021 - 42

An Introduction to Practical Quantum Key Distribution
been some preliminary tests of GEO systems [110]. Additionally, there has been recent work in showing the feasibility of medium earth orbit (MEO) QKD, where it has
been shown that QKD can be conducted with satellites situated in orbits of 7000 km [111], [112]. MEO is less lossy
than GEO, as one might except, while also being more
reliably available than LEO.
The simplest forms of satellite-based QKD have
ground stations each independently conducting QKD protocols with the satellite to establish keys with it [109]. If a
pair of ground stations wish to establish a key between
themselves, they notify the satellite which can securely
communicate a key with each of them individually. For
example, if stations A and B each have communicated
keys Ka and Kb with the satellite, respectively, then the
satellite broadcasts to each station (publicly over an
authenticated but insecure channel) Ka È Kb , allowing A
to recover Kb and B to recover Ka (anyone listening to
Ka È Kb learns neither). In this case, the satellite is acting
as a trusted node.
Trusted node networks [113], [114] are a way to
bridge two quantum end-points allowing for increased distances (especially in fiber links), or for key-establishment
between two parties who do not have current line-of-sight
(for instance, in the satellite case). A trusted node T operates the same protocol as an end-user, establishing two
separate keys with A and B. Call these keys kAT (between
A and T ) and kTB (between trusted node T and user B).
The trusted node then typically broadcasts the bit-string
kAT È kTB (that is, the bit-wise XOR of the strings). This
allows A to compute kTB (which B already knows) thus
giving both parties a shared key. The broadcast of kAT È
kTB does not give an adversary additional information on
kTB (since kAT , being a secret key, is uniformly random
and independent of any adversary's information). The disadvantage, of course, is that T must be trusted (and
secured from outside hacking) as this entity holds the
secret key used by the users.
To agree on these keys, one may choose between
three main link types: uplink, downlink, and retro-reflector simulated downlink. In uplink configurations, a
source at the ground station transmits the quantum states
to a receiver located aboard the satellite, while in downlink configurations the roles are reversed, with the satellite transmitting the states to the ground station. Uplinks,
at the cost of having their detectors be at risk of irradiation from cosmic rays, have the advantage of alleviating
the need to locate a moving quantum signal at, a potentially complicated task, while downlinks suffer fewer
losses than uplinks [115]. Downlinks are the commonly
recommended configuration for satellite QKD, and currently the only demonstrated configuration [116]. A third
possibility is the use of retro-reflectors along with an
uplink to simulate a downlink configuration [117]. In this
scenario, both the source and the receiver are located at
42

the ground station. The ground station transmits a signal
to the retro-reflector aboard the satellite where the signal
is modulated and reflected back to the receiver, thus simulating a downlink without needing to place a transmitter
on the satellite.
A common issue with satellite QKD has been that
their operation has been limited to the night, when the
apparatus are not prohibited by daylight. Recent
work [118], [119], aided by the use of telecommunications
wavelengths and more advanced detectors using integrated photonics [119]-[121], has shown that it is feasible
to achieve long distance quantum communication in both
urban and nonurban environments while experiencing loss
that is comparable to the loss experienced in LEO
systems.
Finally, as a futuristic goal, these kinds of configurations can be combined and chained as parts of larger networks [122]. Satellites can conduct QKD protocols
between themselves to transmit keys from more distant
regions of the network. A large network may consist of
fiber linked metropolitan networks communicating with
both LEO and GEO satellites, those satellites themselves
also communicating with other metropolitan linked networks and other LEO and GEO satellites ultimately resulting in much larger, QKD secured networks than the ones
currently in place. Of course, this is a long-term goal.

LOSS AND KEY-RATE
The loss of a quantum channel, as it turns out, is a fundamental limitation on key-generation rates. Regardless of
the protocol, and even augmented with unlimited two-way
classical communication, it was shown in [123] and[124]
that the secure key generation rate of any QKD protocol is
upper-bounded by a function only the loss in the channel.
If h is the transmittance of the channel, then the so-called
PLOB bound (after the authors of [124] shows the keyrate is proportional to Àlog 2 ð1 À hÞ. This rate may be
overcome using quantum repeater networks, or, more feasible today, through the use of the newly developed TwinField QKD (TF-QKD) protocols [102], [103]. The keyrate for TF-QKD protocols generally scale according to
the square-root of h in theory. In practice, recent experimental results have already broken the PLOB bound [125]
leading to exciting possibilities in scaling QKD infrastructure without necessarily requiring advanced quantum
repeaters.

INTEGRATING QKD TECHNOLOGY
While much research has been done in developing the theory and practice of QKD communication between two
points, there has also been much work done in actually
integrating this technology into current, or future,

IEEE A&E SYSTEMS MAGAZINE

MARCH 2021

IEEE - Aerospace and Electronic Systems - March 2021

Table of Contents for the Digital Edition of IEEE - Aerospace and Electronic Systems - March 2021