Design of ultra-high-capacity optical networks employing space- and frequency-division multiplexing techniques
Loading...
Date
2026-07-10
Authors
Παπαπαύλου, Χαράλαμπος
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
This PhD thesis addresses a fundamental limitation of current optical transport networks, in particular the absence of scalable optical switching architectures capable of supporting the expansion of spectral and spatial dimensions required by next-generation ultra-high-capacity optical networks. While emerging technologies such as Space Division Multiplexing (SDM) and Ultra-Wide-Band (UWB) transmission can offer the necessary degrees of freedom for capacity scaling, existing Wavelength Selective Switch (WSS)-based optical nodes fail to efficiently exploit them due to rigid wavelength-level operation, excessive port scaling, high SWaP-C factor (Size, Weight, Power and Cost) and accumulated filtering impairments.
To address these intrinsic limitations, this thesis proposes a novel hierarchical three-layered Multi-Granular Optical Node (MG-ON) architecture. Unlike conventional designs, the proposed architecture enables native switching across multiple granularities, ranging from entire optical fibers and flexibly defined wavebands, down to individual wavelengths, thereby fundamentally redefining how optical resources are managed in ultra-high-capacity optical networks.
At the core of the proposed architecture lies the flexible Waveband Selective Switch (flex-WBSS), which constitutes a key technological innovation of this thesis, whose fabrication prototype implementation, along with its fundamental operational principles, are presented and thoroughly analyzed in the following chapters.
The flex-WBSS departs from conventional fixed-band wavelength and waveband switching by enabling dynamic spectrum grooming into “flexibly” defined contiguous bands (e.g. S-, C-, L-, with potential future extension to O- and E-bands) across the UWB spectrum (>21 THz). Its novel design, based on Photonic Integrated Circuit (PIC) technology, integrates adaptive Finite Impulse Response (FIR)-based optical filtering, combined with a non-blocking Crossbar Switching (CS) fabric, resulting in a compact, scalable, and energy- and cost-efficient implementation.
A major contribution of this thesis is the end-to-end architectural design of a hierarchical three-layered MG-ON node, where waveband switching is elevated to a primary switching mechanism, bridging the gap between today’s wavelength switching and future full-fiber switching, enabling rapid and scalable network upgrades across both existing and newly deployed infrastructures. This represents a previously missing intermediate layer in optical switching evolution, providing scalable operation without the prohibitive complexity of pure wavelength-based designs.
To rigorously evaluate the proposed architecture, a custom multi-layer optical network simulator is developed, capable of modeling all critical physical-layer impairments while also supporting multiple optical switching policies, including Full-Fiber Switching (FFS), Waveband Switching (WBS), and Wavelength Switching (WS). The obtained results demonstrate substantial improvements in key Quality of Transmission (QoT) metrics such as higher Optical Signal-Noise-Ratio (OSNR) and lower Bit-Error-Rate (BER) compared to conventional WSS-based designs, particularly under ultra-high traffic loads, approaching or exceeding 10 Pb/s. This OSNR improvement is primarily attributed to the reduced number of cascaded filtering stages resulting in lower insertion losses at the physical layer. This overall reduction becomes even more pronounced at the network level (across cascaded nodes), leading to consistently lower accumulated insertion loss across all end-to-end light paths. As a result, signal degradation is significantly mitigated, allowing extended transmission reach with equal or higher OSNR than competing architectures. This, in turn, reduces the number of required active elements, thereby enhancing overall system performance in terms of QoT, power consumption, and cost efficiency.
Furthermore, this thesis develops two complementary frameworks (an analytical and a probabilistic one) for band modeling leveraging static and dynamic traffic matrices respectively, that can be effectively employed in the design of next-generation optical transport networks. The frameworks are employed to evaluate the node’s bandwidth utilization and scaling behaviour, exploring trade-offs among scalability and throughput, as well as component count and complexity. The analysis reveals that the proposed MG-ON architecture achieves substantial improvements over existing counterparts in terms of scalability efficiency, enabling Pb/s-level network throughput while significantly reducing hardware complexity.
Finally, a comprehensive techno-economic and energy efficiency evaluation is conducted across both node and network levels, introducing scalable models developed specifically for this study. The results demonstrate that the proposed PIC-based multi-granular architecture leads to substantial reductions in terms of cost and power consumption.
Overall, this thesis introduces a new architectural and technological foundation for optical switching, positioning the proposed multi-granular, WBSS-based node as a viable pathway toward next-generation optical transport infrastructures.
Description
Keywords
Optical nodes, Ultra wideband transmission, Space division multiplexing, Waveband selective switch