Coherence Big Data Space Division Multiplexing

979-8-3503-6491-0/24/$31.00 ©2024 IEEE Coherence Big Data Space Division Multiplexing Yousef Fazea Department of Computer Sciences and Electrical Engineering, Marshall University, One John Marshall Drive, Huntington, WV 25755, USA, fazeaalnades@marshall.edu Mustafa Muwafak Alobaedy Faculty of Information Technology, City University Malaysia, 46100 Petaling Jaya, Malaysia, Mustafa.theab@city.edu.my Mohamed Alsamman School of Computing, Universiti Utara Malaysia, 06010 Kedah, Malaysia, alsamman@uum.edu.my Fathey Mohammed Sunway Business School, Sunway University, 47500 Selangor, Malaysia, fatheym@sunway.edu.my Abstract— Modal dispersion is well recognized as multimode fiber's primary limitation, leading to connection failures when transmitting huge volumes of data. Numerous central launch conditions have been employed to overcome this obstacle, which had a significant effect on a failure in the connection. These conditions solely induce the primary propagation mode. The potential of utilizing the different light channels in multimode fiber for data centers has not been effectively used. Nevertheless, these fiber modes have the potential to be advantageous for large data applications. At the donut level, beams facilitate the transmission of separate data streams, leading to a higher total bandwidth. The novelty of this work lies in the utilization of the donut beam to optimize computing performance at the infrastructure layer, which, in turn, improves data analytics at the middleware and application layers. By utilizing a central wavelength of 1550.12 nm, achieving a transmission rate of 14 gigabits per second is feasible across a maximum distance of 1200 meters. Keywords—Donut Mode; Vortex lens; Multimode Fiber; Optical Communication I. INTRODUCTION The proliferation of data centers is driven by the ongoing advancement of multi-faceted computing applications. The reason for this is the growing number of internet users and the increasing adoption of cloud computing by businesses for their data infrastructure[1, 2]. Researchers all around the world have been encouraged to explore beyond current speed to change architectures as a result of the continual growth of server, network, and Internet traffic [3-6]. There have been a great number of feasible ways that have been offered to enhance the transmission capacity [7-9]. These approaches include amplitude [10, 11], multiplexing in the polarization [12, 13], wavelength [14-16], and time. Employing different mode excitations such as LG [17], and helical-phase [18]. An emerging technique that has attracted significant interest is known as mode division multiplexing [19-21]. This approach involves modes transmitting their own independent data streams, which ultimately increases the total bandwidth. The utilization of spatial light modulators [22, 23], fiber gratings [24, 25], digital signal processing algorithms [26, 27], modal decomposition algorithms [28], adaptive optics [20], and photonic crystal fiber [29, 30] are some of the techniques that have garnered a lot of attention in recent years. Mode division multiplexing for big data in particular has garnered a lot of attention. In summary, the paper is structured as follows: The subsequent section provides a more comprehensive explanation of the MDM model, which pertains to the donut modes observed in data centers. The paper's conclusion, together with the results and discussion, is offered in Section III and Section IV. II. MODEL SIMULATION Following the steps outlined in Figure 1, the MDM model was built up and simulated using the Optsim 5.2 instrument. It is possible to divide the modal into three distinct components, which are the receiver, the multimode fiber, and the transmitter. All that makes up the transmitter is a single VCSEL operating at a wavelength of 1550.12 nm and being driven by PRBS electrical signals. Modulation that does not return to zero is utilized. According to the definition, the electrical field of the donut mode is: minmax minmax cos( )0 ( , )sin( )0 0 Ll rr r rLl rr r φ α ψφφ ≥  ≤ ≤   =<    < <  (1) the normalization constant is denoted by α. The lowest radius is denoted by r min , and the maximum radius is specified by r max . Within these boundaries, the field is a sinusoidal function of azimuth, where l represents the azimuthal index and L= l|. Simply put, the spot modes are the donut mode with r min =0. During the run duration, the values of x_outer_radius=11um and x_inner_radius=10um were maintained with a beam spot size of 25 microns. The VCSEL is made up of a spatial component that functions as the fundamental donut mode. This component is utilized to govern the launch condition of the donut mode in multimode fiber. The output of the VCSEL will be connected to a vortex lens that is employed for the vortex launch condition. This lens will apply a vortex transformation to the phase front and focus the incoming signal. The launch of the vortex may be controlled by using two parameters, which are the vocal length, which is set to f=8.0mm for a given vortex order of m=4. In the vortex mode, one may describe it as: 2 ( , ) exp ( 2 n r t x yj f π λ  =−   (2) 2 ( , ) exp ( 2 n r t x yjm f π θ λ  =−+   (3) 222 atan( / )r x yandy xθ=+= (4) The variables x and y denote the horizontal and vertical positions, respectively. The symbol λ indicates the length of the signal wave, n represents the refractive index of the material, and f denotes the distance from the lens to the focal point. The parameter m represents the order of a vortex. By using spatial analyzers, it becomes possible to graphically represent both the magnitude and phase of the optical spatial 2024 International Symposium on Networks, Computers and Communications (ISNCC) | 979-8-3503-6491-0/24/$31.00 ©2024 IEEE | DOI: 10.1109/ISNCC62547.2024.10758969 Authorized licensed use limited to: UNIVERSITY TEKNOLOGI MALAYSIA. Downloaded on February 07,2025 at 23:08:04 UTC from IEEE Xplore. Restrictions apply. electrical signal. Subsequently, the signal is conveyed as it traverses through a multimode fiber. Both the power modal coupling and the predicted attenuation values of 1.5 dB/km have been considered. The signal generated by the multimode fiber is examined using a photodetector. The majority of multimode fiber lengths in a typical large data environment are shorter than 500 meters [30]. To fulfill the length requirement for huge data, the mode-division multiplexing (MDM) model that was proposed was able to achieve a distance that ranged from 200 meters to 1200 meters. Following that, the signal is received using a photodetector to be analyzed. To observe the eye diagram and determine the bit-error-rate (BER) of each channel for a variety of multimode fiber (MMF) lengths, the signal is evaluated with a BER tester. III. RESULT AND DISCUSSION The transverse electric field of the donut mode is shown to travel through a vortex lens in Figure 2. This allows the magnitude and phase of the field to be determined. Figure 2(a) illustrates the input incident field of a donut mode with x_outer_radius=11um and x_inner_radius=10um as it travels through a vortex lens with a focal length of 8mm and a vortex order of 4. The output of the special electric is depicted in Figure 2(b), which demonstrates that better results were obtained, and the number of peak intensities can be distinguished. The bit error rate (BER) of the VCSEL array at 1200 meters is depicted in Figure 3, where the BER is equal to 1.83×10 -11 . This is done to investigate the donut mode parameters x_outer_radius and x_inner_radius parameters. When compared to increasing the bit rate that is being communicated, it is important to note that the bit error rate (BER) decreases as the distance between the parties increases. Figure 4 illustrates the impacts of mode delay on the donut mode, namely the outer radius and the inner radius. The xaxis that runs horizontally is referred to as the modal delay (s) and is scaled by a factor of 10 -10 . The vertical y-axis denoted power coupling coefficient. This figure measures how effectively power is being and transmitted distributed in the donut-MDM system. The majority of the data points have a power coupling coefficient that is quite near zero, which indicates that the coupling for those particular modal delays is either low or nonexistent. In addition, there are a few prominent outliers that have power coupling coefficients that are significantly higher than average, which indicates that there is a stronger coupling at those particular modal delays. Figure 1. MDM-based donut mode model (a) (b) Fig. 2: (a) Transverse spatial electric field for photodetector array at λ=1550.12nm before the multimode fiber. (b) Transverse spatial electric field for photodetector array at λ=1550.12nm after the multimode fiber Fig. 3: Bit error rate at a distance of 1200 meters Authorized licensed use limited to: UNIVERSITY TEKNOLOGI MALAYSIA. Downloaded on February 07,2025 at 23:08:04 UTC from IEEE Xplore. Restrictions apply. A power coupling coefficient that is approximately 0.075 is displayed by the peak in the data that is the most apparent, and this value is much greater than the coefficients of the other data points. The power is highly coupled into the mediumorder, which results in a short time delay difference between modes, which ultimately leads to a tiny pulse width. Fig. 5. Illustrate the Delay (sec/m) vs. Degenerate Mode Group (DMG) for OpSig 1. As the red line, which represents delay, exhibits substantial spikes at particular DMG values, particularly in the vicinity of DMG 2 and DMG 6, this inferred that these particular points have a significantly greater delay per unit distance in comparison to other points. The green line, on the other hand, shows an essentially linear increase with DMG, which indicates that there is a direct association between DMG and this metric at this point. A notable deviation can be observed at DMG 6, where there is a drop that is comparable to the spike that can be observed in the red line. Donut x_outer radius and x_inner radius are shown in the eye diagrams in Figure 6, which illustrate the effect of these two parameters on an MMF that is 1200 meters in length. The temporal difference between the propagated modes is growing, and the eye-opening is getting significantly smaller. IV. CONCLUSION Using donut modes generated by VCSEL arrays at a wavelength of 1550.12nm, this study presents the initial numerical investigation of modal characteristics and BER assessment for MDM system. The analysis was carried out using MMF as the transmission medium. Through the MMF, the experimental system was able to reach a data rate of 14 gigabits per second over a distance of 1200 meters, while maintaining BER levels that were within acceptable ranges. The robustness of the proposed method is suggested by the fact that a detailed investigation of relative power distributions and BER performance reveals that optimal signal propagation occurs under these conditions. REFERENCES [1] L. A. Barroso and J. Clidaras, The datacenter as a computer: An introduction to the design of warehouse-scale machines. Springer Nature, 2022. [2] L. a. Tawalbeh, S. Lakkineni, F. Muheidat, U. Bulut, and A. A. Abd El-latif, "Big Data Analytics for Secure Edge-Based Manufacturing Internet of Things (MIoT)," in Secure Edge and Fog Computing Enabled AI for IoT and Smart Cities: Includes selected Papers from International Conference on Advanced Computing & Next- Generation Communication (ICACNGC 2022), 2024: Springer, pp. 135-153. [3] M. Rashmi, D. Girija, and N. Yogeesh, "Exascale Computing: The Next Frontier of High-Performance Computing," Human Cancer Diagnosis and Detection Using Exascale Computing, vol. 279, p. 279, 2024. [4] T. Kumar et al., "Cloud‐based video streaming services: Trends, challenges, and opportunities," CAAI Transactions on Intelligence Technology, 2024. [5] O. Dakkak, Y. Fazea, S. A. Nor, and S. Arif, "Towards accommodating deadline driven jobs on high performance computing platforms in grid computing environment," Journal of Computational Science, vol. 54, p. 101439, 2021. [6] M. Madi, F. Jarghon, Y. Fazea, O. Almomani, and A. Saaidah, "Comparative analysis of classification techniques for network fault management," Turkish Journal of Electrical Engineering and Computer Sciences, vol. 28, no. 3, pp. 1442-1457, 2020. [7] M. Kumari and V. Arya, "Four Laguerre–Gaussian mode-division multiplexed incorporated hybrid XG-PON and FSO Gamma Gamma fading system under severe weather conditions," Optical and Quantum Electronics, vol. 56, no. 1, p. 113, 2024. Fig. 4: The effects of modal delay Fig. 5: Relative group delay at different launch offset. Fig. 6: Eye diagram at a length of 1200 meters Authorized licensed use limited to: UNIVERSITY TEKNOLOGI MALAYSIA. Downloaded on February 07,2025 at 23:08:04 UTC from IEEE Xplore. Restrictions apply. [8] Y. Liu, B. Wang, F. Zhang, and B. Guan, "Highpower and high-quality Gaussian beam in VCSEL via anisotropic modes control," Optics & Laser Technology, vol. 170, p. 110163, 2024. [9] Z. T. Ibraheem, K. K. Ahmed, Y. Fazea, M. Madi, F. Mohammed, and A. Q. Ali, "Boosted PTS method with Mu-Law companding techniques for PAPR reduction in OFDM systems," Wireless Personal Communications, vol. 124, no. 1, pp. 423-436, 2022. [10] M. A. Mosleh, F. Héliot, and R. Tafazolli, "Ergodic Capacity Analysis of Reconfigurable Intelligent Surface Assisted MIMO Systems With a Practical Phase Shift and Amplitude Response," IEEE Transactions on Vehicular Technology, 2024. [11] T. Peng et al., "A novel robustness matching method based on amplitude-frequency characteristic for high power ultrasonic system under variable load," Applied Acoustics, vol. 217, p. 109799, 2024. [12] X. Fu, J. Niu, Y. Huo, S. Yang, and L. Yang, "Polarization-independent reconfigurable WDM- MDM hybrid multiplexer on silicon photonics platform," IEEE Photonics Technology Letters, vol. 35, no. 8, pp. 438-441, 2023. [13] Y. Huo, J. Niu, X. Fu, S. Liu, C. Cheng, and L. Yang, "Active polarization controller based on micro-ring resonators," Optics Letters, vol. 48, no. 21, pp. 5491-5494, 2023. [14] Y. Fazea, A. Amphawan, Y. Al-Gumaei, A. M. Al- Samman, and W. M. Al-Rahmi, "Modes power equalization based-singular value decomposition in mode division multiplexing systems for multihungry bandwidth applications," Optical Fiber Technology, vol. 61, p. 102389, 2021. [15] Y. Fazea, A. Amphawan, and H. Abualrejal, "Wavelength division multiplexing-mode division multiplexing for MMF in access networks," Advanced Science Letters, vol. 23, no. 6, pp. 5448- 5451, 2017. [16] Y. Fazea, "Mode division multiplexing and dense WDM-PON for Fiber-to-the-Home," Optik, vol. 183, pp. 994-998, 2019. [17] Y. Fazea and A. Amphawan, "Mode division multiplexing of helical-phased LG modes in multimode fiber with electronic dispersion compensation," Advanced Science Letters, vol. 23, no. 1, pp. 29-34, 2017. [18] Y. Fazea, "Numerical simulation of helical structure mode-division multiplexing with nonconcentric ring vortices," Optics Communications, vol. 437, pp. 303-311, 2019. [19] Y. Fazea and V. Mezhuyev, "Selective mode excitation techniques for mode-division multiplexing: A critical review," Optical Fiber Technology, vol. 45, pp. 280-288, 2018. [20] F. Wang et al., "OAM mode-division multiplexing IM/DD transmission at 4.32 Tbit/s with a lowcomplexity adaptive-network-based fuzzy inference system nonlinear equalizer," Optics Letters, vol. 49, no. 3, pp. 430-433, 2024. [21] A. E. Willner et al., "Free-space mid-IR communications using wavelength and mode division multiplexing," Optics Communications, vol. 541, p. 129518, 2023. [22] D. Benton, Y. Li, A. Billaud, and A. Ellis, "Spatial Mode Division Multiplexing of Free-Space Optical Communications Using a Pair of Multiplane Light Converters and a Micromirror Array for Turbulence Emulation," in Photonics, 2024, vol. 11, no. 3: MDPI, p. 241. [23] M. Baliyan, A. Shikder, and N. K. Nishchal, "Generation of structured light beams by dual phase modulation with a single spatial light modulator," Physica Scripta, vol. 98, no. 10, p. 105528, 2023. [24] Q. Sun, H. Chen, J. Wang, J. Yang, and H. Jia, "Broadband mode-division (de) multiplexer using nanorod-assisted multimode subwavelength gratings," Optics Communications, vol. 552, p. 130037, 2024. [25] C. Jayanth, S. E. Peter, B. Santhosh, and V. Nandhini, "Design of multiparameter fiber Bragg grating in optical transmission systems wavelength division multiplexing," Journal of Optics, pp. 1-9, 2024. [26] Y. Gao, J. Cui, X. Zhou, Z. Chen, and J. Li, "All- Fiber Low-Modal-Crosstalk Demultiplexers for DSP-Free IM/DD LP-Mode MDM Transmission," in Photonics, 2024, vol. 11, no. 3: MDPI, p. 271. [27] H. Zhou et al., "Turbulence mitigation of four modedivision-multiplexed QPSK channels in a pilotassisted self-coherent free-space optical link using a photodetector array and DSP-based channel demultiplexing," Optics Letters, vol. 49, no. 5, pp. 1209-1212, 2024. [28] J. Li, R. Zhang, F. Wen, F. Yang, B. Wu, and K. Qiu, "Optical Vector-Eigenmode Decomposition Through Deep Learning Networks," Available at SSRN 4726119. [29] Z. Hui, Z. Li, D. Han, and T. Li, "A Novel Mode- Division Multiplexer/Demultiplexer with Ultra- Large Bandwidth and Ultra-Low Insertion Loss Based on Five-Core Photonic Crystal Fiber," 2024. [30] L. He et al., "Wavelength division multiplexing based on the coupling effect of helical edge states in two-dimensional dielectric photonic crystals," Optics Express, vol. 32, no. 7, pp. 11259-11270, 2024. Authorized licensed use limited to: UNIVERSITY TEKNOLOGI MALAYSIA. Downloaded on February 07,2025 at 23:08:04 UTC from IEEE Xplore. Restrictions apply.