5. Channel optimization for enhanced secure key rates

#### 5.1. Experimental setup

The experimental setup for QPSK-based RF-assisted CV-QKD transmission is shown in Figure 9(a). For the transmitter (Alice), a laser with narrow line width is operated at the wavelength of 1550.5 nm with a line width of <50 kHz permitting it to preserve the low phase noise characteristics. The PRBS of length 231–1 is programmed for single wavelength quantum transmission. Resultant 1 GBaud QPSK (four-state) signal is generated after the radio

Figure 9. (a) Experimental set-up for point-to-point QPSK-based quantum transmitter (Alice), quantum channel and quantum receiver (Bob) with hybrid classical 10G traffic and (b) Digital signal processing module for phase noise cancelation (PNC) for quantum signals.

frequency (RF) signals are modulated via an electro-optical I/Q modulator, where RF frequency is kept at 2 GHz. The modulation variance (VA) of the generated quantum signal is optimized by a tunable optical attenuator (TOA). As it is a hybrid classical quantum network, therefore classical 10 Gbits/s QPSK channels are multiplexed at 1531.2, 1571.4, 1591.1 and 1611.2 nm wavelengths. All the classical data channels are optimized at 0.5 mW input power. Whereas, multiplexers and de-multiplexers have 45 dB of isolation between the two adjacent channels, 80 dB isolation between nonadjacent channels and 0.85 dB of insertion loss at 1550 nm. The quantum channel comprises pure silica core fiber (PSCF) with different transmission lengths (maximum = 35 km) and the physical parameters of the fiber under test are enlisted in Section 3.2. The system performance of QKD network is also compared with the SSMF fiber in terms of secure key rates and transmission distance, while Alice and Bob architectures are the same in both the cases.

The coherent receiver (Bob) consists of a 90 optical hybrid, a high optical power handling balanced photo-diodes with 20 GHz bandwidth and a real-time oscilloscope with a 100 GSa/s sample rate and 50 GHz analog bandwidth. We have kept the high power, narrow line width local oscillator at the receiver. It is termed as local local oscillator (LLO). The mean LLO photon level is 1 108 photon per pulse. The line width of the LLO is <10 kHz. After the system calibration at room temperature, the detector efficiency is measured as 0.6, while the electrical noise Vel is 0.85 (in shot noise units). The shot noise variance No is determined with sufficient LLO power to set the balanced detector in the linear detection regime. Shot noise calibration can be performed by shutting down all sources of incoming light or by ceasing the signal optical port on Bob side. The measured No for our setup is ≈ 17 0 mV2. The output signal is processed by the off-line digital signal processing module comprises phase noise cancelation (PNC) algorithm as depicted in Figure 9. The PNC stage has two square operators for in phase and quadrature operators, one addition operator and a digital DC cancelation block assisted by a down-converter. While all the secure key rate measurements are concluded with reconciliation efficiency of 90% for diverse modulation variances and transmission distances [30].

The extended experimental set-up for multiuser optically switched QKD network is shown in Figure 10. A MEMS-based 2 2 switch is incorporated after the quantum channel and demultiplexing, to implement inserted- and by-pass operations. The insertion loss of the switch is measured as 0.8 dB, while the cross talk (XT) is 52 dB, that is, negligible. In a bypass state, the input and output ports are connected to each other and in inserted state, the input and drop ports are connected to each other. In order to recover the classical 10 Gbits/s signals, we have used a standard coherent receiver with built-in DSP module of finite-impulse response filters (FIR) to compensate chromatic dispersion (CD). The forward error correction (FEC) threshold is kept at 3.8 <sup>10</sup><sup>3</sup> BER (bit error ratio).

#### 5.2. Results and discussions

5. Channel optimization for enhanced secure key rates

The experimental setup for QPSK-based RF-assisted CV-QKD transmission is shown in Figure 9(a). For the transmitter (Alice), a laser with narrow line width is operated at the wavelength of 1550.5 nm with a line width of <50 kHz permitting it to preserve the low phase noise characteristics. The PRBS of length 231–1 is programmed for single wavelength quantum transmission. Resultant 1 GBaud QPSK (four-state) signal is generated after the radio

Figure 9. (a) Experimental set-up for point-to-point QPSK-based quantum transmitter (Alice), quantum channel and quantum receiver (Bob) with hybrid classical 10G traffic and (b) Digital signal processing module for phase noise

Figure 8. Optimum system performance and wavelength assignment for hybrid classical-quantum transmission.

5.1. Experimental setup

34 Telecommunication Networks - Trends and Developments

cancelation (PNC) for quantum signals.

We evaluated the average SKRs for diverse modulation variances and transmission distances. While for this investigation, we use two types of quantum channel, that is, SSMF and ultra-low loss PSCF. The results are as shown in Figure 11(a) and (b), respectively. Modulation variance is considered as 0.2, 0.3 and 0.4, while the length of quantum channel is considered upto 35 km

Figure 10. Experimental setup for multiuser optically switched quantum network incorporating 2 2 MEMS add/drop switch for implementing inserted and bypass cases along with hybrid classical 10G traffic.

3.8 103 BER and due to the CWDM channel spacing, we have not seen any inter channel

Figure 12. Experimental secure key rates (SKRs) measurements for multiuser optically switched QKD network with: (a)

Recent Progress in the Quantum-to-the-Home Networks http://dx.doi.org/10.5772/intechopen.80396 37

As we have stated earlier that till date, CV-QKD demonstrations are limited to point-to-point transmission between two distant nodes. For future integration of QKD networks with smart access networks, it is necessary to design a network that can transmit secure keys between multiple parties, hence optical switching techniques may be applied between QKD end-points. Since QKD is very much sensitive to insertion less, noise and cross talk, therefore in our experiment we have investigated a 2 2 MEMS-based switch with measured insertion loss of 0.8 dB, while the crosstalk (XT) is < 52 dB, that is, negligible and switching time is 20 ms. This is a two position device, that is, insertion and by-pass state as shown in Figure 10, that is commonly termed as optical add-drop multiplexer. In the by-pass operation, the input and output ports are connected to each other, that is, Alice is connected to Bob-1. On the other hand, in insertion operation, the input and drop ports are connected to each other, that is, Alice is connected to Bob-2, while at the same time, we can connect the add and output port for different set of secure keys. We have achieved 5.98 Mbit/s of secure key rates for almost both of the inserted and by-pass state at 20 km transmission distance, as in Figure 12. We certainly believe that in multiuser QKD network the optically switched key rates can further be improved with efficient splicing/coupling with same matching fiber, since in our case the MEMS 2 2 switch has 9/125 μm single mode fiber. Nevertheless, this key rate is much higher than the recently reported results of 4.75 Mbit/s for 1.5 dB attenuation (corresponding to 7.5 km quantum channel). The results discussed in this section are helpful to develop quantum secure routers that require high secure key rates, switching speed and low loss QKD optical switch.

In this chapter, we have given the theoretical design of a QTTH network and estimated the potential of using the commercially available equipment to generate the secret quantum keys.

cross talk between the classical and weak quantum channels.

by-pass operation and (b) inserted state operation.

6. Summary of the chapter

maximum due to the limitations of resources in the laboratory. The maximum SKRs of 8.65 Mbit/s can be obtained with ultra-low loss PSCF-based quantum channel at 20 km transmission distance, as shown in Figure 11(b). It can be seen that for the same transmission distance with SSMF-based quantum channel (Figure 11(a)), the average SKRs are reduced to 5.9 Mbit/s. It is evident from the results that the ultra-low loss PSCF-based quantum link can give you enhanced transmission distance with much improved key rates. It is worth mentioning here that since we are talking about the low baud rate signals and access networking distances, that is, 20–30 km, therefore PSCF fiber is performing better than SMF. It will be necessary to have dispersion mitigation module along with phase noise cancelation module as an integral part of Bob over much longer distances due to higher dispersion factor of PSCF fiber. It is further noticed that all the classical 10G channels are detected below the FEC threshold level, that is,

Figure 11. Experimental secure key rates (SKRs) measurements for diverse modulation variance values with respect to transmission distance for: (a) standard single mode fiber (SSMF) and (b) low loss pure silica core fiber (PSCF). The detector efficiency is 60% and reconciliation efficiency is 90%.

Figure 12. Experimental secure key rates (SKRs) measurements for multiuser optically switched QKD network with: (a) by-pass operation and (b) inserted state operation.

3.8 103 BER and due to the CWDM channel spacing, we have not seen any inter channel cross talk between the classical and weak quantum channels.

As we have stated earlier that till date, CV-QKD demonstrations are limited to point-to-point transmission between two distant nodes. For future integration of QKD networks with smart access networks, it is necessary to design a network that can transmit secure keys between multiple parties, hence optical switching techniques may be applied between QKD end-points. Since QKD is very much sensitive to insertion less, noise and cross talk, therefore in our experiment we have investigated a 2 2 MEMS-based switch with measured insertion loss of 0.8 dB, while the crosstalk (XT) is < 52 dB, that is, negligible and switching time is 20 ms. This is a two position device, that is, insertion and by-pass state as shown in Figure 10, that is commonly termed as optical add-drop multiplexer. In the by-pass operation, the input and output ports are connected to each other, that is, Alice is connected to Bob-1. On the other hand, in insertion operation, the input and drop ports are connected to each other, that is, Alice is connected to Bob-2, while at the same time, we can connect the add and output port for different set of secure keys. We have achieved 5.98 Mbit/s of secure key rates for almost both of the inserted and by-pass state at 20 km transmission distance, as in Figure 12. We certainly believe that in multiuser QKD network the optically switched key rates can further be improved with efficient splicing/coupling with same matching fiber, since in our case the MEMS 2 2 switch has 9/125 μm single mode fiber. Nevertheless, this key rate is much higher than the recently reported results of 4.75 Mbit/s for 1.5 dB attenuation (corresponding to 7.5 km quantum channel). The results discussed in this section are helpful to develop quantum secure routers that require high secure key rates, switching speed and low loss QKD optical switch.
