**4. Techniques and structure in QKD: challenges and trends**

The techniques and structures used in QKD context involve the different set-ups, operation rules and devices that perform a particular protocol. Therefore, the first step is choosing the quantum protocol and next, the general structure can be proposed and implemented. In particular, the structure consists of optical source, optical detector, digital processing unit (the challenges and trends that have already been mentioned) among other specific devices connected together in order to perform a complete QKD system. On the other hand, the techniques are the novel operational rules in order to enhance the complete performance of the QKD system. Each protocol mentioned was proofed, first, using a particular technique and structure, these can be can be found and analyzed in the references listed. However, many improves to each protocol have been proposed for QKD systems implemented in FSO.

For example, the atmospheric turbulence is an important problem for QKD systems based on FSO links. In order to mitigate the degraded performance of the secret key rate for QKD systems based on BB84 protocol, an optimization technique was proposed based on an adaptive optical power transmission considering the random irradiance fluctuation [28]. In the same context, a novel encoder technique was proposed for the classical channel in QKD-FSO systems based on adaptively encoder gain according to atmospheric turbulence levels [29], The results show that the secret key rate remains constant for a region of turbulence levels and imposes the need of a high-end DP&Comm subsystem in order to extend the operating region. In addition, many structures and techniques used in conventional classical optical communication systems have been adapted to QKD-FSO systems. In particular, Multi-Input-Multi-Output (MIMO) and Wavelength Division Multiplexing (WDM) are suitable options used in order to increase the capacity of free space channel based on Orthogonal Angular Momentum (OAM) modulation [30]. Among the structures and techniques necessary to implement a QKD-FSO system are the subsystems used in order to pointing, acquisition and tracking the two parties (Alice and Bob, represented by satellites and ground stations). In this case, pointing systems used in satellites have reached from 0.6 μrad to 3 μrad pointing capability [31, 32].

## **4.1 Challenges and trends**

In general, the structures and techniques allow to improve the performance of a QKD-FSO system. Therefore, the design of techniques and high-end structures allows to support in a better way the actual QKD system proposals. In fact, the principal challenges are related with the optimization and improving of the secondary subsystems of a QKD-FSO systems (i.e. secondary subsystems are not mentioned in detail in this chapter, such as telescope, mechanical structures, access multiplexing techniques, among others). Finally, the QKD-FSO system trends related with the structures and techniques are: maximize the channel capacity, increase the distance link and secret key rate, increase the power consumption efficiency in order to support long-time missions, improve the thermal control and isolation, among others.

In addition, novel encoding technique for classical channel has been proposed in order to increase the secret key rate at QKD-FSO links. **Figure 4** shows a diagram proposed based on an adaptive LDPC (Low-Density Parity-Check Codes) encoder in order to countermeasure the effect caused by the dynamical atmospheric turbulence [29].

*Quantum Cryptography in Advanced Networks*

to improve the performance of particular QKD systems.

The challenges present in the QKD protocols are related with the performance parameters of the QKD systems. In particular, although each protocol uses different security principle and quantum states, the important issues are increasing the security level, secret key rate and distance link between Alice and Bob in presence of Eve system. In fact, while a particular protocol presents a high security level and particular secret key rate for short distance links, other protocol presents the same security level and secret key rate for long distance links. However, as was mentioned, a QKD protocol requires the other subsystems, thus, a hypothetically complicated protocol imposes a strict and detailed design, that is, the experimental set-up is not simple. Therefore, the tendency of the protocols refers to proposing novel QKD protocols that allow to easily implement them in optical commercial networks, while the performance parameters remain constant or improved. In addition, a high dimension protocol is proposed in order to increase the photon information capacity when the photon rate is restrained. This protocol is based on entangled photon pairs that allow information to be transmitted using an extremely large alphabet [27]. Now, each QKD protocol has been theoretically described, however, free space and atmospheric channels impose important trade-off that determines the suitable protocol. In particular, BB84 protocol has been optimized for FSO links affected by atmospheric turbulence improving the secret key rate up to over 20% [28].

**3.1 Challenges and trends**

While BB84 protocol uses 4 orthogonal states, B92 only uses 2 non-orthogonal states. Therefore, the different quantum states (i.e. orthogonal and non-orthogonal) used in BB84 and B92 protocols impose a trade-off regarding the final secret key rate generated by Alice, Bob and the attacks performed by Eve [20]. Since the BB84 protocol is extremely vulnerable to Photon Number Splitting attacks, the SARG04 protocol was proposed, which uses 4 non-orthogonal quantum states; however, the final secret key rate is also affected [21]. Additionally, there exists the E91 protocol based on Einstein, Podolsky and Rosen (EPR) paradox that uses entangled quantum states generated either by Alice, Bob or a trusted third party [22]. Later, the BBM92 protocol was proposed which implies EPR pairs, that is, entangled photon pairs. This protocol can be described as the BB84-EPR protocol [23]. Until now, the protocols mentioned are based on State of Polarization (SOP), DV framework and general stages such as: raw key exchange, key sifting and privacy amplification, that is, all the protocols have the same stages in order to generate the final quantum key. On the other hand, QKD protocols based on CV variables are also suitable, such as COW protocol (Coherent One-Way), which is based on an amplitude encoded sequence of weak coherent pulse with the same phase for each particular time slot. In particular, different time slots have several optical pulses (related to an optical power average) and, occasionally, decoy sequences are sent in order to hinder the eavesdropped process [24]. Due to the different quantum states and encoding scheme used, this protocol is so-called distributed-phase-reference (DPR), in fact, there are many protocols in the same category such as the differential-phase-shift (DPS), which uses different phases but the amplitude remains constant. Therefore, interferometric techniques are required in the receiver [25]. All the DPR protocols perform joint measurements on subsequent signals. Actually, GG02 protocol is present in many commercial equipment. In general, this protocol is based on random distributions of coherent or squeezed states and modulates either the phase or amplitude of a quantum state and uses coherent detection in Bob's side [26]. Finally, each protocol mentioned has a particular security principle, be it the Heisenberg uncertainty or quantum entanglement. Although there exist novel protocols that change the security principle in order

**32**

**Figure 4.**

*Block diagram of the simulation-experimental set-up of the QKD system proposed with dynamical encoder for different atmospheric turbulence levels. Own figure and presented in [29].*

Here, *I*(*A:B*) is the mutual information that Alice and Bob shared, and the maximum information shared for Alice and Eve is *S*(*A:E*). In this case, βgg is the reconciliation efficiency. In other hand, α is the classical channel efficiency that is based on the encoder capacity (related to the amount the erroneous bits that are detected and corrected). In this scenario, the dynamical atmospheric turbulence is represented by Rayleigh or Gamma-Gamma (GG) density probability functions, *Pe*(α,σ*<sup>R</sup>* 2 )*c* and *Pe* (α,α*gg*, <sup>β</sup>*gg*)*<sup>c</sup>* , respectively, where <sup>σ</sup>*<sup>R</sup>* 2 represents the Rytov variance related with the atmospheric turbulence, *αgg* and *βgg* are the effective numbers of large-scale and small-scale for GG function, respectively. Basically, Alice and Bob monitored the dynamical atmospheric turbulence calculating the error probabilities and modifying LDPC encoder capacity used by them.
