Optimizing Multi-Band Optical Network Design: A Layered Approach for Engineering and Education

Abstract

The sixth generation of mobile networks (6G) presents increasing complexity that challenges traditional analysis and performance evaluation methods, necessitating more structured approaches for both research and educational purposes. This study introduces a layered methodology that classifies physical layer impairments, such as amplified spontaneous emission (ASE) noise and fiber nonlinearities into sequential layers. The approach enables independent assessment of individual impairment contributions to overall system performance, facilitating more accurate evaluation of signal quality metrics, including signal-to-noise-ratio (SNR) and optical signal-to-noise-plus-interference ratio (OSNIR) across multiple spectral bands. By implementing this step-by-step analysis framework, researchers can better understand the cumulative impact of various transmission effects, while students can gain progressive insight into complex optical communication principles, making this approach serve dual purposes as both an effective research tool for system optimization and a pedagogical instrument that enhances engineering education. The effectiveness of the methodology is demonstrated through the performance evaluation of a system employing five spectral bands (E, S1, S2, C, and L) under various operating conditions.

1. Introduction

As the sixth generation of mobile networks (6G) will satisfy the user demands for higher data rates compared to the fifth generation of mobile networks (5G) [1], the optical transmission systems will need to extend their capabilities by increasing both the number of spectral and spatial elements [2]. During this extension, network complexity increases and the process of understanding and isolating the major physical effects and parameters of the physical layer that affect the system’s performance becomes cumbersome. As a consequence, network design is becoming more difficult, requiring more sophisticated modeling approaches and advanced optimization algorithms to manage the interplay between the various physical layer impairments that maximize transmission performance under various operating conditions. Moreover, this complexity poses significant pedagogical challenges for engineering educators, as traditional teaching methodologies are inadequate for conveying the interdependencies between multiple physical phenomena [3], requiring innovative educational approaches that can effectively demonstrate system-level interactions and provide students with the conceptual frameworks necessary to navigate multidisciplinary problem-solving in 6G technological environment.

To address these challenges, this work introduces a layered approach for optical network design: a four-layer methodology that decomposes complex multi-band optical systems into manageable, sequential layers, each focusing on specific physical impairments. This approach draws inspiration from established telecommunications models like the OSI and TCP/IP frameworks, where complex systems are broken down into distinct functional layers for better analysis and understanding. In this methodology, physical layer impairments, including receiver noises, amplified spontaneous emission (ASE) noise, nonlinear interference (NLI), and stimulated Raman scattering (SRS), are systematically organized into four sequential layers.

The layered framework allows us to assess the transmission performance in terms of signal-to-noise ratio (SNR) and optical signal-to-noise-plus-interference ratio (OSNIR) [4]. Our main pedagogical goal is, by introducing this methodology, to provide students with a better and deeper understanding of modern optical networks, regarding their capabilities and limitations and to help them develop a layered and scalable approach to assess and optimize the performance of optical networks. Moreover, this approach will give the opportunity to professors to easily address the individual needs of each student that might be present in a classroom and enable them to learn in their own way, pace, and time, as the model promotes engagement and self-directed learning. It also gives students freedom to analyze each network layer separately, with precision and effective problem-solving. Overall, we aim to give a solid framework for everyone interested, might it be a student, a researcher, or an engineer, for tackling each layer with confidence and understanding more in depth the impact that every physical impairment has on an optical multi-band system.

The candidate framework offers several advantages: (a) reduced complexity, as it segregates the system into manageable layers, each with a specific function, (b) simplified teaching and learning, as it makes 6G network-related concepts easier to understand and teach, (c) better understanding of the main causes of performance degradation, as the physical layer effects are studied both separately and in combination, (d) local optimization, as it enables per layer optimization leading eventually to global optimization, (e) cost efficiency, as it allows a modular upgrade at a layer level, leading to a lower CapEx and OpEx for a given performance, and (f) multiple levels of complexity per layer, as the impact of a physical layer impairment can be estimated using various methods, e.g., analytical, numerical, or machine learning (ML) [5,6] based on the sought level of accuracy.

In prior art, the layered approach has been adopted in a variety of domains, like software design, to support the management of alternative software architectures [7], telecommunications, for aiding fault diagnosis [8], security, for Internet legal analysis [9], and edge computing, for supporting smart grid Internet of Things (IoT) implementations [10]. While these works provide a clear demonstration of the adaptability and versatility of the layered approach, none has applied it in the field of optical networks, nor integrated it into a telecommunications curriculum. The current work aims to be the first to introduce and systematically apply the layered approach to optical network design, both as a robust framework for analyzing the impact of the dominant physical layer impairments, individually and in combination, and also as an innovative pedagogical tool to deepen and enhance university students’ understanding of the modeling, analysis, and performance evaluation of these complex systems.

Overall, the main novelties of this work are as follows: (i) to the best of our knowledge, this is the first work that proposes a layered approach for the design of optical networks where it introduces four layers, each dealing with different effects. This approach allows for the focused study and modeling of each impairment (first three layers) and also for the examination of the combined impact of all effects (fourth layer); (ii) educational innovation: This is the first work that introduces the layered approach in the field of optical networks aiming to improve and simplify the educational process by allowing students to better comprehend complex network behaviors by decomposing them into comprehensible layers, fostering improved understanding, active learning and self-acquired skills.

2. Related Work on Layered Approach

This section presents the layered approach and its implementation into the teaching process. Subsequently, the use of the layered approach in optical network design is tackled, analyzing how it can aid in enhancing students’ understanding as well as benefit the evaluation and optimization design of optical networks.

2.1. Layered Approach for Improving the Teaching Process

The layered approach in education is a differentiated instructional model that is designed to focus on all the diverse needs, abilities, and learning styles that the students in the same classroom can have. The most common implementation of the layered approach is the layered curriculum [11]. Its main advantage is that, for each student independently, any misconceptions they might have can be easily spotted and addressed, therefore avoiding big gaps in the knowledge that each student gains by the course. This way students can feel more included in the learning process, feel more motivated to learn, and emphasize more in their autonomous learning. Research shows that the layered approach improves academic achievement, retention of learning, and the development of autonomous and self-directed learning. By structuring the learning method into progressively challenging layers, the students, being surer of themselves in every learning step they take, are able to tackle more problems with better skills.

Table 1. Utilization of the layered approach in prior art.

Ref.	Domain	Details
[7]	Software Engineering, Design Environments	Improves the development and management of alternative software designs since there are limitations of file-based systems in representing alternatives and integrating operational environment.
[8]	Telecommunication Networks, Fault Diagnosis	Automates fault diagnosis in dynamic telecommunication networks since the complexity and dynamism of modern networks make static models inadequate.
[9]	Internet Law, Policy Analysis	Provides a conceptual framework for analyzing Internet legal and policy issues by reflecting its layered technical architecture, since traditional legal analysis often fails to account for the Internet’s multi-layered structure, leading to incomplete or misdirected policy responses.
[10]	Smart Grid, IoT	Leverages IoT for upgrading the power grid into a smart grid because it needs bidirectional communication, automation, and management in modern energy networks.
[12]	Telecommunications Education, 5G/F5G Networks	Introduces a unified, layered understanding of both wireless (5G) and fixed (F5G) network generations for engineering students and educators, to address the (i) fragmented teaching of telecom topics that leads to knowledge segmentation and misconceptions and (ii) lack of unified perception of fixed and wireless networks.
[13]	Grid Computing, Optical Networks	Supports computationally demanding applications through Grid-over-OBS architecture due to the need for efficient routing and delay minimization in grid applications.
[14]	Educational Sciences, Language Teaching	Determines whether layered curriculum enhances academic achievement in English for 9th grade students because of the persistently low English proficiency among students despite curriculum reforms. Needs more effective, individualized teaching methods.
[15]	Language Teaching, Student Perceptions	Explores student perspectives on differentiated, student-centered learning since traditional English lessons can be monotonous and lack the engagement.
[16]	Teacher Education, Reflective Thinking	Enhances reflective thinking and self-directed learning readiness in pre-service teachers because the usual teacher training often lacks focus on reflective practice and learner autonomy.
[17]	Science Education, Learning Styles	Addresses diverse learning styles and improves engagement in science education since the standard approaches tend to overlook individual learning preferences and hinder effective engagement.
[18]	Telecommunications Policy, Regulatory Models	Clarifies the original intent and utility of the SMC layered policy model and addresses misconceptions and criticisms in regulatory frameworks, because the misinterpretations and competing models have led to confusion and policy misapplications regarding the layered approach in telecom regulation.
[19]	Pharmacy Education, Clinical Training	Addresses the growing need for quality experiential training in pharmacy by integrating multiple learners into clinical settings because of the increased number of pharmacy students’ and residents’ strains preceptor capacity and risks compromising learning quality.
[20]	Educational Sciences, Language Teaching	Adapts language teaching to individual differences and improves student achievement and attitudes due to the ineffectiveness of traditional methods in addressing differences and learning styles.
[21]	Vocational English Education	Addresses the uneven English proficiency of vocational students and the ineffectiveness of traditional uniform teaching, because the traditional way of teaching may fail to promote English proficiency due to diverse student backgrounds and abilities.
[22]	Educational Technology, Science Education	Determines if differentiated instruction and technology can enhance comprehension and motivation since traditional teaching methods fail to engage all students and do not address learning preferences or technology integration.
[23]	Educational Sciences, Curriculum Design	Addresses individual differences and fosters higher-order thinking through differentiated instruction since the traditional, uniform instruction fails to account for diverse student abilities, interests and learning styles.
[24]	Educational Sciences, Learning Retention	Improves academic achievement and retention through student-centered, differentiated learning, because the conventional approaches do not sufficiently promote responsibility, critical thinking, or retention.

tabulates prior art that uses the layered approach to address some already established problems within different domains and sub-domains.

In reference to Table 1, if we examine each work separately, we conclude that each one has their advantages and areas for improvement. These are outlined in a detailed list below:

Ref. [7]: This work claims that through the layered approach, collaborative software design is enhanced, and redundancy is reduced by integrating both code and related documents. However, that comes at a relative “cost” of the systems becoming more complex and more difficult to learn, especially in larger projects.

Ref. [8]: With the implementation of this model, autonomy, adaptability, and learning with large networks are achieved, but on the downside, there is a greater complexity and a need for more training data, and the computations can be too complex.

Ref. [9]: As this layered model brings to light certain issues and precise jurisdictional definitions and gives an analytical approach to internet policy and law, it still may fail to address some real-world complexities; confusion can arise between the layers, and, of course, for the proper implementation, technical and legal expertise is required.

Ref. [10]: The implementation of the proposed approach provides a more holistic view of the challenges and the opportunities that smart grids have to offer. However, the system is highly complex, with issues of interoperability and security that are not yet solved.

Ref. [12]: With the teaching methodology used in this work, the educational gap that exists between wireless and fixed networks is closed effectively, while a more holistic mindset is developed, which allows students to align themselves with the day-to-day industry developments. However, it is limited by a brief course duration, and as F5G technology continues to progress with each passing day and becomes more standard, further curriculum adaptation will more than likely become a necessity.

Ref. [13]: While the proposed approach enables more to the point resource management and reduces delays, further research on data aggregation and retransmission is needed.

Ref. [14]: The layered curriculum as stated by this work helps in boosting academic achievement and motivation, as well as in increasing the feeling of responsibility through differentiated learning. However, because of the complexity of this method’s implementation, it needs to be thoroughly explained and planned, and the teachers must be properly trained, while they are also responsible for each learner’s readiness.

Ref. [15]: While students are responsible for choosing the level and type of work that suits them best, which encourages teamwork, it has the disadvantage that not every student may engage at the same level, as some may choose less demanding and difficult tasks over the more challenging ones.

Ref. [16]: With the implementation of such a curriculum, reflective thinking and autonomy are developed; however, bigger and more flexible learning spaces are needed, as well as a higher percentage of readiness for self-directed learning.

Ref. [17]: The teaching of the lesson becomes more engaging and diverse learning styles and needs are easier to address, though a certain adaptation to the teaching method of the appointed teacher is needed and gender-specific differences are highlighted.

Ref. [18]: This methodology offers a systematic and technically neutral policy analysis tool, which helps with more targeted regulation and also cultivates innovation at the content/application levels. However, the downside of this is that it risks oversimplification and can sometimes struggle to keep up to date with the rapid technological changes, resulting in incorrect readings of the conditions of the real-world telecom market.

Ref. [19]: According to the author, by using the layered learning model in this work, an expansion in the educational capacity is achieved, promoting leadership and engagement, but requiring a well-developed planning process to be implemented, clearly defined roles, and continuous feedback.

Ref. [20]: By implementing the layered curriculum, we obtain higher motivation, critical thinking, and responsibility, but how effective this method can be depends on the planning and teaching expertise to adapt to varying circumstances.

Ref. [21]: By using the said methodology, the results are ongoing progress for all and deeper engagement by the students, as it accommodates many diverse student abilities. However, as addressed before, its success percentage is dependent on well-supported and adaptable teachers, and even then, it may not be effective in addressing some deeper disparities that exist among the students.

Ref. [22]: The layered teaching methodology, aided by means of technology, can help improve understanding and address student diversity, but even then, student motivation may not always be significantly affected; for it to be successful, several technological instruments are needed, which are most of the times quite “pricey”.

Ref. [23]: This approach, once again, develops autonomy, authentic learning, and deeper engagement across various learning environments, but it requires expert teachers that can adapt this approach to the preexistent curricula.

Ref. [24]: By using the method of the layered curriculum, critical thinking is cultivated, as well as responsibility and retention, albeit with the need for an active participation from the students and the adaptation of learners to self-directed activities.

From the analysis of the aforementioned literature, it is clear that the layered approach is applied across technology, policy, and education to enhance integration, adaptability, engagement, and analytical depth. It can efficiently support design, network management, policy analysis, smart grid implementation, and differentiated learning, among others. The main benefits of these works include improved collaboration, targeted regulation, and fostering critical thinking and autonomy. However, challenges such as high complexity, steep learning curves, the need for expert adaptation, technological demands, and dependence on well-prepared educators or planners still need to be overcome. In the next sub-section, we present the literature and the advantages of the candidate layered approach for optical network design.

2.2. Layered Approach for Optimizing the Optical Network Design

The proposed approach classifies the complex physical layer phenomena that can degrade the performance of an optical multi-band system into four layers as illustrated in Figure 1. This methodology provides a better understanding of each physical layer effect and also of the overall optimization of the optical system, where the interactions between the different impairments arise along the transmission path. By layering the physical layer impairments sequentially, we can determine the impact of each major impairment in a separate layer, allowing for investigation of the dominant source(s) of performance degradation and facilitating the development of customized mitigation strategies that specifically target each impairment type.

In addition, the candidate approach can further enhance students’ understanding: first, about each effect individually, and second, regarding the interactions between them. This can also aid greatly in identifying each student’s misconceptions, allowing for targeted interventions to address them. This approach builds a solid understanding before addressing the more complex real-world systems, minimizing errors during the design, implementation, and deployment of a real network.

Next, the modular structure allows more targeted mitigation strategies for each impairment and for each band in a multi-band optical network, thus supporting better overall management across different spectral bands. Finally, the layered approach is inspired by well-known models in telecommunications (e.g., TCP/IP, OSI) [25,26], thus enhancing its pedagogical value and practical relevance for network design and troubleshooting. By taking a look at the OSI model, we can see that it is built on seven layers (physical, data link, network, transport, session, presentation, application), with each one performing certain functions. Owing to its structure, it can isolate the different processes as well as the impact of physical phenomena and protocols on the overall network. The TCP/IP model in comparison consolidates the functions of several OSI layers, as it is built with five functional layers (physical, data link, network, transport, application).

Furthermore, as is evident from [27], when a similar methodology was used in a telecommunications lab course at ASPETE, in the academic year of 2019–2020, where due to the COVID-19 pandemic the courses were being carried out online, it aided in better identifying the students’ typical errors and misconceptions. More specifically, the study of [27] explains how, by layering the lab curriculum into different exercises—starting from the analytical estimations for the separate impact of the various phenomena and ending with the design of an optical link using numerical simulations, considering the joint impact of multiple effects—students’ understanding was significantly aided; as it seems, they had the time to focus their attention every time on a single new concept, and the exercises were stepping stones for the students to progress to the next one with the knowledge of the previous ones.

Overall, the proposed four-layer classification approach not only enables the targeted identification of performance degradation sources and customized mitigation strategies per band but also enhances educational outcomes by facilitating individual impairment analysis before progressing to complex interactions.

3. Layered Approach for Optical Network Design

In this section, we detail the proposed layered approach for the design of optical multi-band networks.

3.1. Overview

The layered approach in telecommunications breaks down complicated communication systems into separate layers, each investigating different effects and network elements, e.g., transceivers, fibers, and amplifiers. This makes it easier to understand, build, fix, and upgrade optical networks, as each effect can be treated separately from the others. Regarding the educational aspect, a layered model, like the OSI or TCP/IP model mentioned previously, is very useful in the classroom, because it helps students understand how data communication works step by step. By looking at each layer on its own, students can build a solid conceptual base that is necessary for a deeper understanding and application in real-world systems.

3.2. Analytical Modeling of Multi-Band Transmission

Analytical modeling of multi-band transmission is a crucial process for understanding the fundamental mechanisms that affect signal quality and network performance. In practical optical links, multiple spectral bands are often exploited to maximize capacity, but the strength of each physical layer impairment is different in each band. The proposed approach addresses this complexity by assessing the impact of the various performance degradation mechanisms individually in the first three layers, enabling clearer isolation, characterization, and evaluation of their individual impact. Then, in the fourth layer, the joint impact of all effects is estimated, allowing additional optimizations using all effects together. This methodology facilitates both tractable analysis and physical insight by decomposing the total transmission impairment into additive or interacting sub-effects and by composing all impairments together. It also helps with a better understanding of what affects the optical channel, in what way, and how detrimental it can be.

The different physical layer effects are classified into four layers as follows:

• Layer 1: The receiver noises, thermal and shot noise are investigated.
• Layer 2: The contribution of the ASE noise created by the optical amplification is calculated.
• Layer 3: The NLI caused by the Kerr effect is studied.
• Layer 4: The impact of all the aforementioned effects, in addition to the SRS, is considered and included in a single performance evaluation metric.

This layered structure enables a clear decomposition of the complex physical layer effects and helps us evaluate the SNR, OSNIR, and Bit Error Rate (BER) for various operational cases.

The proposed methodology has a direct correspondence with the Marzano–Kendall taxonomy, which reaches up to its fourth level. The four layers of the proposed model mirror the first four cognitive levels outlined by Marzano and Kendall: retrieval, comprehension, analysis, and knowledge utilization [28], but not in a one-to-one correspondence. By layering and isolating each physical layer degradation, it enables learners and researchers to first recall and identify (retrieval), next obtain a better understanding of the significance and mechanisms (comprehension), then compare and decompose their effects on system performance (analysis), (Layer 1 to 3 of the proposed approach) and finally use this layered understanding to optimize system design and evaluation (knowledge utilization) [28] (Layer 4 of the proposed approach). The close alignment with the aforementioned taxonomy supports both a better and in-depth scientific evaluation and enhances the pedagogical value of the proposed work.

The following sections develop each layer in detail, offering a way of calculating every degradation with accuracy by using analytical expressions and finally highlighting their cumulative impact on optical multi-band transmission systems. It is worth mentioning that the proposed layered approach is not confined only to the analytical formalism presented below but can also include other methods for performance evaluation, such as analytical, e.g., the GN-model and its variants [29,30,31,32]; numerical, e.g., the Split Step Fourier Method (SSFM) [33]; ML-based [34]; even a combination [6,35,36]; and experimental methods [37,38], each differing in terms of accuracy, estimation time, computational complexity, the reality gap they exhibit relative to real system behavior, and their level of explainability.

The highest accuracy is obtained by experimental methods since they incorporate the impact of the physical phenomena that are usually neglected by theoretical methods. In addition, ML methods, when trained with real data, can attain high accuracy; however, typically, this accuracy is confined to a relatively small number of operational cases, as it is costly to collect extensive datasets from experimental systems. Further, ML methods suffer from limited generalizability beyond training conditions and reduced explainability. Numerical methods like SSFM offer high accuracy for complex scenarios but demand substantial computational resources and longer processing times while suffering from limited explainability.

In contrast, analytical methods offer superior computational efficiency, mathematical transparency, and deeper insight into system behavior [5]. They enable rapid design space exploration, parameter sensitivity analysis, and closed-form relationships between system parameters and performance metrics, enhancing student comprehension of complex optical phenomena. The transparent nature of analytical expressions allows educators to demonstrate cause-and-effect relationships between physical parameters and system performance, facilitating step-by-step learning that aligns with established educational taxonomies, such as [28]. In addition, students can easily visualize how individual parameters influence overall system behavior, fostering deeper conceptual understanding and enabling self-directed learning at personalized paces. Overall, analytical approaches exhibit a larger reality gap compared to experimental ones; however, their combination of speed, explainability, reasonable accuracy, and exceptional pedagogical value makes them well-suited for the engineering design and educational applications targeted by the proposed layered framework.

3.2.1. Receiver Noises

In the first layer, the impact of the receiver noises is considered. In an optical network, receiver noises encompass all of the unwanted electrical fluctuations that are generated internally by the receiver’s components, such as low-noise amplifiers, mixers, and filters. Receiver noises primarily consist of thermal and shot noise, and these negatively affect the system’s ability to detect and decode the received signal. Minimizing receiver noises is critical for improving signal fidelity, maintaining data integrity, and ensuring efficient use of spectrum, particularly in bandwidth-limited or low-power communication environments. In a core or metropolitan area optical network, the thermal and shot noise are not the dominant sources of noise, but in the access part of the network, where optical amplification is usually not required, they are two of the main physical layer impairments.

To calculate the shot and thermal noise variances, we exploit the following expressions

$${\sigma }_{shot}^2=2q\left({\mathrm{R}}_e\cdot P_r+I_{dark}\right)B_e$$

(1)

$${\sigma }_{thermal}^2={\displaystyle \frac{4kT{B}_e}{R}}$$

(2)

where q denotes the elementary charge equal to 1.6 × 10⁻¹⁹ C; R_e is the responsivity of the photodetector, which is wavelength-dependent as follows $R_{e }={\displaystyle \frac{(\eta \cdot \lambda \cdot {10}^6)}{1.23985}}$ and measured in A/W; P_r is the received power; I_dark is the dark current, which is typically too small and can be neglected, k is the Boltzmann constant equal to 1.38 × 10⁻²³ (J/K); T is the absolute temperature equal to 300 K; R is the load resistor equal to 50 Ohm; and B_e is the bandwidth of the electrical filter, which in our case is equal to the baud rate.

3.2.2. ASE Noise

In the second layer, the power of the ASE noise is estimated. ASE noise is the “unfortunate” result of optical amplification due to spontaneous emission. This noise arises when spontaneously emitted photons within the gain medium are themselves amplified, producing a wideband background signal that can interfere with the transmission of the desired optical signal. Unlike receiver noise, ASE noise is broadband and accumulates after every amplification stage. Because of that, it directly affects the degradation of OSNIR in an optical path and becomes critical in long-haul fiber links, as a result of the multiple amplifiers used in these links.

We calculate the power of the ASE noise by using the following equation:

$$P_{ASE}={\displaystyle \sum _{i=1}^{Ns}\left[hf\left(N{F}_i\cdot G_i-1\right)B_o\right]}$$

(3)

where N_S denotes the number of fiber—xDoped Fiber Amplifier (xDFA) spans, h represents Planck’s constant (6.62 × 10⁻³⁴ J·s), which is fundamental in determining the energy of individual photons, B_o denotes the optical bandwidth over which the ASE noise is measured, equal in this work to 50 GHz, while G_i and NF_i denote the ith amplifier’s Gain and noise figure, respectively.

3.2.3. Nonlinear Interference

In the third layer, the most deleterious phenomenon is the NLI. NLI is important when the power of the light injected into the fiber is high, which makes the fiber’s refractive index depend on the optical power. This mainly leads to the creation of unwanted frequencies, which can make it difficult to correctly extract information from the received signal. In wavelength division multiplexing (WDM) systems, NLI is even more important than in single channel transmission, as the various channels interfere through this mechanism and a large number of unwanted frequencies are generated. As a consequence, the effective management of NLI is essential to maintain sufficient performance and minimize interference. The following closed-form expression can estimate the power of NLI [4]:

$$\begin{array}{l}{P}_{NLI}={\displaystyle \frac{32}{27}}{\displaystyle \frac{\gamma {(\lambda )}^2{L}_{eff}^2{P}_{ch}^3{N}_s^{2}c}{{\lambda }^2{B}^{2}D(\lambda )\sqrt{z_1}}}\left(1+{\displaystyle \frac{4e^{-\alpha (\lambda )L}}{{\left(1-e^{-\alpha (\lambda )L}\right)}^2}}\right)\left(\mathrm{asinh}\left({\displaystyle \frac{\pi {\lambda }^{2}D(\lambda )B^2}{8c}}{N}_{ch}^{{\displaystyle \frac{2B}{B+GB}}}\sqrt{z_2}\right)-{\displaystyle \frac{5}{3}}\Phi \log \left(N_{ch}^{{\displaystyle \frac{B}{B+GB}}}\right)\right)\\ -{\displaystyle \frac{32}{27}}{\displaystyle \frac{\gamma {(\lambda )}^2{L}_{eff}^2{P}_{ch}^3{N}_s^{2}c}{{\lambda }^2{B}^{2}D(\lambda )\sqrt{z_1+12L^2}}}\left({\displaystyle \frac{4e^{-\alpha (\lambda )L}}{{\left(1-e^{-\alpha (\lambda )L}\right)}^2}}\right)\left(\mathrm{asinh}\left({\displaystyle \frac{\pi {\lambda }^{2}D(\lambda )B^2}{8c}}{N}_{ch}^{{\displaystyle \frac{2B}{B+GB}}}\sqrt{z_2+12L^2}\right)-{\displaystyle \frac{5}{3}}\Phi \log \left(N_{ch}^{{\displaystyle \frac{B}{B+GB}}}\right)\right)\end{array}$$

(4)

where $z_1={\left({\displaystyle \frac{2}{a\left(\lambda \right)}}\right)}^2+2L^2\left({N_s}^2-1\right)/{\left({\displaystyle \sum _{k=x_1}^{x_2}\frac{1}{1+{\left(2k\pi /\left(a\left(\lambda \right)L\right)\right)}^2}}\right)}^2$, $x_1=-{\displaystyle \frac{{\lambda }^2{B}^{2}D\left(\lambda \right)L{N_{ch}}^2}{16c}}$, $x_2={\displaystyle \frac{{\lambda }^2{B}^{2}D\left(\lambda \right)L{N_{ch}}^2}{2c}}$, $z_2={\left({\displaystyle \frac{2}{a\left(\lambda \right)}}\right)}^2+2L^2\left({N_s}^2-1\right)$, γ(λ) is the nonlinear fiber coefficient, D(λ) is the local dispersion, and a(λ) is the fiber attenuation parameter. These three parameters are wavelength dependent. Next, L denotes the span length, N_ch is the number of transmitted channels in each band, B is the channel bandwidth, GB is the channel spacing, c is the speed of light in vacuum (3 × 10⁸ m/s), L_eff denotes the effective fiber length, and finally Φ is a modulation format dependent parameter, which in our case equals to 1 as the employed modulation format is Polarization Multiplexing (PM)-QPSK. For 16QAM and 64QAM, Φ equals 17/25 and 13/21, respectively.

3.2.4. Stimulated Raman Scattering

In the fourth layer of our study, the impact of SRS is quantified as well as the interplay between SRS, ASE noise, and NLI. SRS is a nonlinear optical process that occurs when high-intensity light propagates through an optical fiber. In essence, it transfers energy from channels with higher frequencies (shorter wavelengths) to channels with lower frequencies (longer wavelengths). SRS can cause a significant power loss in lower bands, e.g., O, E, and S, and unwanted “amplification” in C and L bands. This effect changes the power distribution of the signal across the spectrum and can significantly degrade the performance of all bands if not properly engineered. In modern high-capacity fiber-optic networks, power management across all channels is important for reducing the impact of SRS. The SRS-induced loss/gain consists of two components as in [39]:

$$G_{SRS}=G_{SRS,\le 15THz}+G_{SRS,>15THz}$$

(5)

The first term accounts for the interaction of channels with frequency differences up to 15 THz and can be estimated by

$$G_{SRS,\le 15\mathrm{THz}}=P_{\mathrm{tot},SRS}{\displaystyle \frac{e^{\frac{g^{\prime }\cdot B\cdot L_{\mathrm{eff}}}{2A_{\mathrm{eff}}}(j-1)P_{\mathrm{tot},SRS}}}{{\displaystyle \sum _m\left[P_{m,0}{e}^{\frac{g^{\prime }\cdot B\cdot L_{\mathrm{eff}}}{2A_{\mathrm{eff}}}(m-1)P_{\mathrm{tot},SRS}}\right]}}}$$

(6)

where g′ is the Raman gain slope equal to 4.9 × 10⁻²⁷ m/(W∙Hz), A_eff is the effective cross-sectional area of the fiber, and P_m,0 is the power of the mth interfering channel at fiber input. The term P_tot,SRS sums the power of channels that interact within the 15 THz spectrum, and B is the spectral distance of the interfering channels. The second term in (5) is for channels with a spectral distance > 15 THz and is given in (7), where k is the index of a channel with a spectral distance > 15 THz (k > m) from channel j, and P_j and P_k are their launch power at fiber input, respectively. The SRS gain in this spectral region is calculated with the aid of the quantity G_SRS,peak, which is the maximum Gain/Loss value in (7) between channels with index j and m. With the aid of (6)–(7), the SRS-induced power exchange for a spectrum up to 35 THz can be estimated.

$$G_{SRS, >15 \mathrm{THz}}={\displaystyle \sum _k\left[\frac{P_k{G}_{SRS, peak}}{0.85P_j}\left(0.49-0.08\cdot Log\left(\frac{|k-j|B-15}{0.05}+1\right)\right)\right]}$$

(7)

Finally, given that P_i is the channel power, and P_ASE,i and P_NLI,i denote the powers of the ASE and NLI accumulation in the ith channel, respectively, the P_ASE,i, when the SRS is present, is given by [6]:

$$P_{ASE}={\displaystyle \sum _{i=1}^{Ns}\left[hf\left(N{F}_i\cdot G_i-1\right)B_o{\displaystyle \prod _{r=i+1}^{Ns}{G}_{SRS,r}}\right]}$$

(8)

3.2.5. Metrics for Performance Evaluation

To evaluate the performance of an optical multi-band transmission system, we exploit the following metrics that include the impact of the aforementioned effects.

• Signal-to-noise Ratio (SNR): This metric is fundamental for assessing the quality of the received signal. It provides the theoretical performance limit for each spectral band before we include additional effects, such as ASE noise and NLI. The SNR, including the received power P and the impact of receiver noises, can be estimated through the following expression

$$SNR\approx Q_{factor}^2,\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}{Q}_{factor}={\displaystyle \frac{{\mathrm{R}}_e\cdot \mathrm{P}}{\sqrt{{\sigma }_{th}^2+{\sigma }_{sh,0}^2}+\sqrt{{\sigma }_{th}^2+{\sigma }_{sh,1}^2}}}$$

(9)

where the shot noise is calculated for different power levels, e.g., for bit “0” and for bit “1” when direct detection is considered. In cases where the SNR must account for additional noise sources, e.g., ASE noise and fiber nonlinearities, the expressions of [40,41] need to be employed.

• Optical Signal-to-Noise-Plus-Interference Ratio (OSNIR): OSNIR is the extension of SNR, as it incorporates the effects at the optical part, in particular, the fiber nonlinearities, the SRS, and the ASE noises. It is used to measure the accumulation of every major source of performance degradation in the optical system.

$$OSNIR={\displaystyle \frac{P_i}{P_{ASE,i}+P_{NLI,i}}}$$

(10)

• Bit Error Rate (BER): BER is the measurement of the accuracy of the received bits in the transmission, and it can be directly related to the OSNIR for different modulation formats as follows:

$$BE{R}_{QPSK}={\displaystyle \frac{1}{2}}erfc\left(\sqrt{{\displaystyle \frac{OSNIR}{2}}}\right),\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}BE{R}_{16\hspace{0.17em}QAM}={\displaystyle \frac{3}{8}}erfc\left(\sqrt{{\displaystyle \frac{OSNIR}{10}}}\right),\hspace{0.17em}\hspace{0.17em}\hspace{0.17em}BE{R}_{64\hspace{0.17em}QAM}={\displaystyle \frac{7}{24}}erfc\left(\sqrt{{\displaystyle \frac{OSNIR}{42}}}\right)$$

(11)

where the erfc(.) denotes the complementary error function.

Table 2. System parameters and constants used in our study.

	E Band	S1 Band	S2 Band	C Band	L Band
λ (nm)	1416.5	1466.7	1496.7	1546.9	1594.6
α (dB/km)	0.280	0.246	0.229	0.211	0.210
D (ps/nm/km)	8.63	12.06	13.97	16.96	19.60
γ (1/W/km)	1.65	1.50	1.44	1.32	1.24
Aeff (μm2)	70	74	76	80	83
NF (dB)	6	5.5	5.5	5.5	6
Nch	101	69	66	86	105
f (THz)	211.64	204.271	199.06	193.7	188.06
B (GHz)	50
g (m/(W·GHz))	4.9 × 10−18
c (m/s)	3 × 108
λ0 (nm)	1310
h (J·s)	6.62 × 10−34
L (km)	50
k (J/K)	1.38 × 10−23
T (K)	300
q (C)	1.6 × 10−19
R (Ohm)	50
η	0.8

includes all the constants and parameters considered in the current study.

4. Results

In this section, the results of our study are presented, layer by layer, with the aim of demonstrating how every degradation affects optical transmission, to what extent, and how it can be optimized. To gain a better understanding and a clearer picture of how everything comes into play, the performance is evaluated by iterating for different values of the channel power (P) and the number of fiber spans (Ns). The optical power is one of the most critical parameters, affecting the effects in all four layers, while the number of fiber spans plays an important role in the accumulation of ASE noise and nonlinear effects.

4.1. Layer 1—Receiver Noises

Figure 2 illustrates the SNR for different values of the received optical power. In this figure, the impact of only shot and thermal noise is included, defining the theoretical minimum received power (P) to achieve a target performance in terms of SNR in a multi-band optical system. This figure demonstrates that the SNR increases linearly as a function of received channel power (P) across all spectral bands (E, S1, S2, C, L), showing a difference between the channels in different bands, mainly due to the fact that the responsivity is wavelength dependent. The analysis indicates that a received optical power of more than −20 dBm can ensure sufficient performance in terms of SNR. By having our first layer refer only to the receiver noises, a clear reference point for subsequent layer analysis is established, namely, before introducing the complexity of ASE noise, NLI effects, and SRS in the next layers. Such a stepwise approach is essential for building intuition and for the evaluation of the physical impairments an ultra-wideband optical network can face.

4.2. Layer 2—ASE Noise

This layer estimates the impact of ASE noise on four different distances, from 300 to 1200 km, with an incremental step of 300 km. Figure 3 illustrates the P/P_ASE versus P for the five different sub-bands. We can observe that as the number of spans increases, the P/P_ASE ratio decreases linearly for all the bands, as ASE noise accumulates linearly with the number of spans since the signal traverses more amplifiers. By comparing the various bands in the plots in Figure 3, it is revealed that the L band achieves the highest P/P_ASE values, while the E band has the lowest ones, which can be attributed to the higher fiber attenuation in the E band and consequently to the larger required amplification gain that leads to greater ASE noise. The linear dependence of P/P_ASE on launch power, while excluding nonlinearities and SRS, makes it clear that ASE noise is the dominant physical layer constraint, compared with shot and thermal noise, indicating a minimum of about −8 dBm in E and S bands and about −11 dBm in C and L bands in order to overcome its impact.

4.3. Layer 3—Fiber Nonlinearities

In this layer, the impact of NLI is quantified. Figure 4 illustrates the P/P_NLI for the same transmission distances as the previous sub-section. As is evident, the system performance decreases as the launch power increases. This is expected from the expression of Equation (4), where the NLI power is proportional to the cube of the optical power, and as a result, the OSNIR degrades at higher power levels, e.g., after 0 dBm. In addition, for a given launch power, the P/P_NLI ratio decreases as the number of spans increases, reflecting the cumulative buildup of the NLI over long transmission distances. We can also observe that the power of NLI is higher at lower wavelengths, as the chromatic dispersion parameter D(λ) is increasing with the wavelength and the P_NLI is inversely proportional to D(λ), as presented in Equation (4).

4.4. Layer 4—All Effects

In the fourth and final layer, the impact of all major physical layer impairments, which are ASE noise, NLI, and SRS, is considered, and the OSNIR performance across all the spectral bands (E, S1, S2, C, L) is shaped by their interaction. Figure 5 illustrates OSNIR values for different power levels designating also the required OSNIR values that lead to a target BER of 10⁻³ for PM-QPSK, PM-16QAM, and PM-64QAM, according to Equation (11). From this figure, it can be clearly seen that OSNIR evolves as a function of channel launch power (P) for increasing transmission distances (6, 12, 18, and 24 spans). When low power is launched, the OSNIR increases with power, because the system is in the ASE-limited regime, while as the launched power increases further, under the specific system configuration examined, the system reaches its optimal power region at approximately −5 dBm for E and S bands and around 0 dBm for C and L bands. After these values, the OSNIR decreases, since the system reaches the nonlinear-limited regime, where the NLI and SRS dominate. These optimal operating points are dependent on the fiber parameters, amplifier characteristics, and transmission distance and may vary in different network configurations.

Another important observation is the strong impact of SRS, after a specific power, e.g., −5 dBm, which redistributes optical power from lower-wavelength bands (E, S1, S2) toward higher-wavelength ones (C, L), making the lower bands operate in ASE-limited regime, significantly degrading their OSNIR, leading to a more than 5 dB difference, compared with the higher bands. Also, the cumulative impact of ASE and NLI leads to a lower OSNIR in longer transmission distances as they accumulate with the number of fiber spans. It is also worth mentioning that it is possible to exploit power allocation methods, such as two-zone OSNIR optimization (higher OSNIR in C and L bands with less than 0.5 dB variance and lower OSNIR in E and S bands with less than 0.5 variance), to achieve a flat OSNIR performance across all bands with less than 0.5 dB variance and weighted merit function approaches to achieve differentiated connectivity schemes, maximize network throughput, and enhance wavelength routing flexibility based on the high-level objectives set by the network designers [42].

5. Discussion and Recommendations

The results of our study show that the layered approach provides a clear and effective method for analyzing and optimizing optical multi-band transmission systems by introducing the main physical impairments in different segments and isolating them: first, the receiver noises; next, the ASE noise; then, the NLI; and finally, the SRS. This approach helps in evaluating the separate impact of each effect on the system’s performance. Our findings demonstrate that receiver noises and ASE noise dominate at low launch powers and shorter distances, while NLI and SRS increase significantly as the launch power and distance grow. The candidate layered analysis highlights the necessity of careful power management to maximize OSNIR, especially due to the power transfer between bands through SRS.

While this work demonstrates the effectiveness of the layered approach using E, S, C, and L bands as a representative case study, the analytical framework is inherently scalable to additional spectral bands, power allocations, and modulation formats through parameter modifications (e.g., Φ parameter for various QAM formats). Regarding cross-layer interactions, the proposed framework captures them through Layer 4, which estimates the cumulative and interactive effects of all physical impairments, including ASE noise, NLI, and SRS, as demonstrated in the OSNIR calculations and power redistribution analysis. The proposed methodology provides a structured approach for understanding how lower-layer effects propagate and interact in the overall performance assessment.

From an educational perspective, the structure of the layered approach helps better support understanding and gradual skill development. It allows students to build their understanding “from the ground up”, by learning from the basic noise analysis to the more complex interplay of nonlinear and inter-band effects. This makes the approach suitable for telecommunications curricula, and it can be integrated into lab exercises and projects, where students can progress at their own pace in learning, aiding them to better prepare themselves for real-world engineering. On the other hand, applying the layered approach in network design enables targeting certain parts of the network and focusing on their optimization. Future work is recommended to enhance this framework by including additional performance degradation factors in the four layers and by integrating it with digital learning tools, further benefiting its educational value.

While the layered approach provides design clarity, multi-band optical networks face significant interoperability challenges, including vendor-specific equipment compatibility and varying amplifier technologies across spectral bands. Security vulnerabilities include fiber tapping, high-power jamming, and cross-band interference attacks exploiting nonlinear effects. Mitigation strategies should incorporate standardized APIs for cross-vendor compatibility, unified power management protocols, physical layer encryption, and continuous optical performance monitoring with anomaly detection. Future framework extensions must integrate security-by-design principles and interoperability standards while preserving the educational transparency of the systematic layer-by-layer methodology.

6. Conclusions

In conclusion, the layered approach provides greater insight into the physical phenomena that affect an optical multi-band transmission system. By segregating them into different layers, the receiver noises, ASE noise, NLI, and SRS, it is illustrated more clearly how each impairment behaves within realistic network scenarios. The results show that by carefully monitoring the launch power per band, we can optimize the OSNIR, especially as the transmission distance and the number of transmitted channels increase. Moreover, the method used by the layered approach does not only aid in network design, but also in the teaching process, by enhancing the conceptual clarity of the scenario. All in all, given the double benefit we obtain from the layered approach, one can say that it can be considered an essential tool for both researchers and educators alike in tackling the modern-day complexities of optical communications.

This work could be enhanced further by implementing this approach in a classroom, and after finishing all the layers and the whole lesson, the students could be asked for their feedback as to how they liked the form the lesson took, how would they compare it to other lessons, if it helped them understand better the effects of each layer, and how they affect the system as a whole, as well as the possible drawbacks this lesson form may have. The proposed method can also be implemented and offered as an open-access tool, for students, educators, researchers, and engineers to use as they see fit and be able to compute each layer and see how each degradation responds and affects the whole optical system. Finally, the logic that exists behind the layered approach, as it is already been established, can also be applied in various telecommunication sub-domains, like wireless communications, ML applications in optical networks, educational content optimization, and software-defined networking, where the systematic decomposition of complex systems into manageable layers enhances both understanding and implementation efficiency across diverse technological domains.