Secure Data Processing Technology of Distribution Network OPGW Line with Edge Computing

Promoted by information technology and scalable information systems, the network design and communication method of optical fiber composite overhead ground wire (OPGW) have been in great progress recently. As the overhead transmission line has strict requirements on the outer diameter and weight of OPGW, it is of vital importance to perform the physical-layer secure data processing for the distribution network OPGW line with edge computing. To this end, we examine a physical-layer secure distribution network OPGW with edge computing in this article, where there exists one transmitter S, one receiver D, one authorized legitimate monitor LM, and an interfering node I. To better analyze the system performance, we firstly give the definition of the system outage probability, based on the secure data rate. Then, we evaluate the system performance for the distribution network OPGW, by deriving analytical outage probability of secure data processing, to facilitate the system performance evaluation of secure data processing in the entire SNR regime. Finally, we demonstrate some simulation results to validate the analytical results on the physical-layer secure distribution network OPGW line with edge computing.


Introduction
Promoted by information technology and scalable information systems [1][2][3], the network design and communication method of optical fiber composite overhead ground wire (OPGW) have been in great progress recently, where the overhead transmission line has strict requirements on the outer diameter and weight of OPGW [4,5].There are two feasible directions to increase the power communication capacity.One direction is to reduce the size of optical fiber and increase the capacity of fiber core in the optical fiber cable.The second direction is to reduce the loss of optical fiber and improve the communication capacity and transmission rate of single fiber.In order to adapt to the development of power communication, the research on two types of small size single-mode optical fiber and ultra-low loss G.654.E optical fiber has become a hot spot.In 2015, Corning released SMF-28 Ultra 200 optical fiber with an outer diameter of 245uM reduced to 200um.In 2019, Prysmian released BendBrightXS 180uM bend insensitive optical fiber, which is 50% smaller than traditional optical fiber, and compatible with G652 and G657 optical fiber standard.The OPGW used for ultra-high voltage is of stranded structure.
The OPGW made of 652.D optical fiber increases the number of optical fiber cores from 30 cores and 24 cores to 48 cores and 36 cores, respectively [6].OPGW has excellent performance and provides a solution for power communication capacity expansion.However, the bending resistance of conventional G.652.D optical fiber is poor, and the micro bending loss after the coating is thinned deteriorates.Hence, it is necessary to study and design the G.652.D optical fiber, which has a better compatibility with a larger mode field bending resistant fiber.At the same time, the bending resistance of the fiber after the coating diameter should be reduced, which can further enhance the system performance of OPGW network.
The G.654.E optical fiber for land use can reduce the attenuation while increasing the effective coverage, reduce the nonlinear effect, allow a larger input power, and achieve a longer distance for communication [7].It has greater application advantages in remote areas and harsh environment with ultra high vacuum (UHV) communication.The transmission distance of single cross optical without relay is about 460km, and using G652 can no longer meet the requirements of longer distance, which imposes a serious challenge on the traditional communication system of ultra-long distance optical fiber.To solve this challenge, 8-core 130 with uM2 G.654 has been configured for the E optical fiber, and strict requirements on the G.654.E optical fiber has been put forward.After cabling, the average two-way average of single fiber at 1550nm wavelength is smaller than 0.165dB/km, which is far lower than the requirement in ITU-TG.654.E, where 1550nm attenuation is smaller than 0.23dB/km.However, lower attenuation means higher manufacturing cost, which needs to be combined with practical scenarios and new technologies of communication and computing.
Edge computing provides an effective solution to monitor the IoT based networks including the OPGW networks and 5G networks, where an intensive data analysis is involved [8,9].Due to limited computing resources such as limited computing frequency and limited energy supply, one single computing node may not calculate all of the computing tasks alone, where the edge computing has been proposed to accelerate the computing efficiently [10,11].In this area, many existing works have been performed to investigate the application of edge computing into the IoT based networks including the OPGW networks and 5G networks [12,13].For example, the computing latency can be effectively reduced by using the edge computing techniques, when the transmission quality of offloading the tasks is in good quality.In addition, the computing energy consumption can be also reduced effectively, when the offloading channel is in good condition, providing a good trade off between the computing and communication.Hence, it becomes an important task to investigate the effect of edge computing on the IoT based OPGW networks, especially the monitoring performance.
From the above literature review, we can find that it is of vital importance to perform the physical-layer secure data processing for the distribution network OPGW line with edge computing, where physical-layer security provides an effective solution to monitor the Internet of Things (IoT) based networks including the OPGW networks and 5G networks, in which an intensive data analysis is involved.Due to limited computing resources such as limited computing frequency and limited energy supply, one single computing node may not calculate all of the computing tasks alone, and accordingly edge computing is utilized for the transmission of OPGW networks.Hence, in this article, we examine a physical-layer secure distribution network OPGW with edge computing, where there exists a transmitter S, a receiver D, an authorized legitimate monitor LM, and an interfering node I. To better analyze the system performance, the definition of the system outage probability is firstly given based on the secure data rate difference between the main and monitoring links.Then, we evaluate the system performance for the distribution network OPGW, by deriving an analytical expression for the outage probability of secure data processing.Finally, the simulation results are demonstrated to show the validity of the analytical results on the physicallayer secure distribution network OPGW line with edge computing.The results in this work can provide some importance references for the development of information technology and scalable information systems. .

System model of the secure data processing of distribution network OPGW line with edge computing
In this paper, we consider a secure distribution network OPGW line with edge computing, where there is one transmitter S, one receiver D which can act as the computing access node, one interference node I, and one authorized legitimate monitor LM.The transmitter S sends some intensive tasks to the edge node D via the wireless link for computing services, while the LM can monitor the data transmission in the network.In addition, the interference node I interferes with the wireless transmission of both data link and monitoring link.Specifically, let h ∼ CN (0, β 1 ) and g ∼ CN (0, β 2 ) denote the channel parameters from S to D and S to LM, respectively.In addition, h I ∼ CN (0, β 3 ) and g I ∼ CN (0, β 4 ) are the channel parameters from I to D and I to LM, respectively.For the considered system, the received signal-to-noise ratios (SNRs) at D and LM are given by [14][15][16] SNR D = P S |h| 2 where P S is the transmit power at S and P I is the interfering power at I. From ( 1) and ( 2), we can find that a larger P I will deteriorate the received SNRs at D and LM, which leads to a poor data transmission performance as well as the monitoring performance.
According to (3), the legitimate monitoring outage happens if the data rate of the monitoring link is smaller than the transmission data rate.Hence, we can further write (3) as [19][20][21] where we use From ( 6) and ( 7), ( 5) can be rewritten as [25,26] P out = +∞ 0 A 1+

P S P I
x −1 We can further derive P out as where we use Then, we will discuss P out for different relationships between a and b.Specifically, if a = b, we can have, For the other case of a b, we can have By summarizing the analytical results in ( 15) and ( 17), one can readily perform the system performance evaluation of the secure data processing in the considered network.

Simulations on the secure data processing of distribution network OPGW line with edge computing
In this part, in order to verify the analytic results derived above, simulation results and analytical results  are presented for the secure data processing of distribution network OPGW line with edge computing.In particular, the impact of parameters Y th , β 1 and β 2 on the performance of the entire monitoring network is analyzed and simulated in the subsequent several figures.If we do not give the specific values, P I , P S , and Y th are equal to 1W, 2W and 0, respectively, and the average channel gains β 1 , β 2 , β 3 and β 4 are all set to unity.
Fig. 2 illustrates the system outage probability with β 1 = 1 or 2 versus the threshold Y th , where the threshold Y th varies from 0bps to 1bps.As shown in Fig. 2, when β 1 = 1 or 2, the system outage probability increases with the increase of the threshold Y th .For example, when β 1 = 1, the system outage probability with Y th = 0 is about 0.5, while that with Y th = 1 is about 0.7.When β 1 = 2, the system outage probability with Y th = 0 is about 0.39, while that with Y th = 1 is about 0.57.Moreover, we can also find that the average channel gain β 1 also affects the legitimate monitoring outage probability.In particular, the result with β 1 = 2 is smaller than that with β 1 = 1 when Y th ranges in the interval of [0, 1].In addition, the analytical P out is almost equal to the simulation P out .This validate the derived closed-form expressions for the legitimate monitoring outage probability.For example, the analytical outage probability with β 1 = 2 and Y th = 0.2 is 0.4240, while the simulated outage probability with β 1 = 2 and Y th = 0.2 is 0.4239, where the difference is only 0.0001.Such results verify the derived outage probability for the considered system.Fig. 3 depicts the variation of the system outage probability with β 2 = 1 or 2 versus the average channel gain β 1 , where the average channel gain β 1 varies in [1,10].As shown in Fig. 2, when β 2 = 1 or 2, the outage probability decreases with the increase of the channel gain of β 2 .For example, the system outage probability with β 1 = 1 and β 2 = 1 is about 0.50, while the system outage probability with β 1 = 10 and β 2 = 1 is about 0.17.The system outage probability with β 1 = 1 and β 2 = 2 is about 0.61, while the system outage probability with β 1 = 10 and β 2 = 2 is about 0.25.This is because that increasing β 1 can help increase the corresponding SNR D , which in turn leads to a lower legitimate monitoring outage probability.Moreover, we can also find that the channel gain β 2 also affects the legitimate monitoring outage probability.In particular, the system outage probability with β 2 = 2 is lower than that with β 2 = 1.In addition, the analytical P out can match the simulation P out well.This also validate the derived closed-form expressions for the system outage probability.Fig. 4 presents the impact of the average channel gain β 2 on the system outage probability with β 1 = 1 or 2, where the average channel gain β 2 changes in [1,10].As shown in Fig. 4, when β 1 = 1 or 2, the system outage probability increases with the increasing channel gain β 2 .In addition, increasing β 1 can effectively increase the corresponding SNR LM , which in turn leads to a higher outage probability.For example, the system outage probability with β 1 = 1 and β 2 = 1 is about 0.50, while the system outage probability with β 1 = 2 and β 2 = 1 is about 0.39.The system outage probability with β 1 = 1 and β 2 = 10 is about 0.83, while the system outage probability with β 1 = 2 and β 2 = 10 is about 0.75.Moreover, we can also find that the average channel gain β 1 also affects the legitimate monitoring outage probability.Specifically, the value of P out with β 1 = 1 is lower than that with β 2 = 2.In addition, the analytical P out fits well with the simulated P out .This further   validates the derived closed-form expressions for the system outage probability of the considered network.

Conclusions
In this paper, we investigate the secure data processing of distribution network OPGW line with edge computing, where the intensive tasks may be involved to perform the data analysis on the network maintenance.To assist the the system performance analysis, we firstly gave the definition of the system outage probability based on the secure data rate difference between the main and monitoring links.Then, we evaluated the system performance for the distribution network OPGW, by deriving an analytical expression for the outage probability of secure data processing.Finally, the simulation results were demonstrated to show the validity of the analytical results on the physical-layer secure distribution network OPGW line with edge computing.The results in this work could provide some importance

Figure 1 .
Figure 1.The system model of the secure data processing of distribution network OPGW line with edge computing, where there is one transmitter S, one destination D, one legitimate monitor LM and one interference node I.

Figure 2 .
Figure 2. Impact of threshold Y th on the system outage probability.

Figure 3 .
Figure 3. Impact of the average channel gain β 1 on the system outage probability.

Figure 4 .
Figure 4. Impact of the average channel gain β 2 on the system outage probability.

Table 1
Numerical outage probability versus the threshold Y th .

Table 2
Numerical outage probability versus the average channel gain β 1 .

Table 3
Numerical outage probability versus the channel gain β 2 .