Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor | Department of Electrical and Electronic Engineering | en_US |
| dc.contributor.advisor | Lau, Pak Tao Alan (EEE) | en_US |
| dc.creator | He, Yan | - |
| dc.identifier.uri | https://theses.lib.polyu.edu.hk/handle/200/14535 | - |
| dc.language | English | en_US |
| dc.publisher | Hong Kong Polytechnic University | en_US |
| dc.rights | All rights reserved | en_US |
| dc.title | Quality of transmission modeling and estimation for live commercial optical fiber networks | en_US |
| dcterms.abstract | Optical fiber networks form the critical backbone infrastructure for high-speed, high-bandwidth transmission supporting modern AI and Internet applications. Recently, open and disaggregated optical networks have gained popularity, transitioning from legacy single-vendor closed systems to multi-vendor disaggregated systems. This architecture allows network operators to choose best-in-class devices, enabling customized network functionality and management planes. However, its multi-vendor characteristics make physical layer modeling much harder due to the accuracies of system parameter characterizations and the implementation principle of the devices can vary with different vendors. To design and manage disaggregated networks efficiently, accurate knowledge of physical layer impairments (PLI), or Quality of Transmission (QoT), on each lightpath is essential. Given that inaccurate input parameters still degrade the accuracy of the QoT estimation, in this regard, this thesis proposed novel methods to improve the accuracy of QoT estimation for different types of commercial networks with disaggregated architecture. | en_US |
| dcterms.abstract | This thesis first studied a large-scale long-haul optical network and proposed to refine the signal power measurement and amplifier gain profile used in the QoT model. This method was further extended by adding simultaneous learning of EDFA NFs, additional biases for different OMS and frequencies. The monitored data has a period of 8 months and represents the largest-scale disaggregated network study for QoT estimation with 1,478,354 data samples. The std of estimation error using the proposed techniques can be reduced from 0.6337 to 0.2377dB. In addition to long-haul, this thesis also studied metro data center interconnect (DCI) networks where the data-rate of TRX varies from 200Gbps to 800Gbps and are more disaggregated with multiple vendors. We first trained an ANN model to learn inherent transponder effects for accurate QoT estimations. Besides, this thesis also proposed input refinement method for analytical SNR model with a specific focus on one type of 400Gb/s TRx. The refinement method was based on experimentally characterizing the input-dependent TRx SNR penalty and input-dependent NFs of multi-vendor multi-type EDFAs. The RMSE of QoT estimation using our method can be reduced from 0.662dB to 0.287dB. This thesis also proposed accurate polynomial-based amplifier gain profile characterization model for multi-vendor multi-type commercial EDFAs. The accuracy of greenfield QoT estimation enabled by the power prediction and EDFA gain profile model is comparable to that of the brownfield network, demonstrating that even at the pre-deployment stage of the network, reliable QoT estimation can be achieved. Finally, this thesis provided an open-source dataset extracted from a commercial live network where the optical power, bit error ratio (BER) and etc., are made publicly available on GitHub, providing a benchmark for comparing different QoT models. | en_US |
| dcterms.abstract | In summary, this thesis presented improved brownfield and greenfield QoT estimations that enable accurate and fast assessment of signal impairments applicable in different commercial network architectures. These models and methods facilitate automated network design, planning and operation, enabling operators to optimize capacity utilization, reduce unnecessary overprovisioning and avoid manual interventions or rework. | en_US |
| dcterms.extent | xvi, 147 pages : color illustrations | en_US |
| dcterms.isPartOf | PolyU Electronic Theses | en_US |
| dcterms.issued | 2026 | en_US |
| dcterms.educationalLevel | Ph.D. | en_US |
| dcterms.educationalLevel | All Doctorate | en_US |
| dcterms.LCSH | Optical fiber communication | en_US |
| dcterms.LCSH | Telecommunication systems | en_US |
| dcterms.LCSH | Neural networks (Computer science) | en_US |
| dcterms.LCSH | Hong Kong Polytechnic University -- Dissertations | en_US |
| dcterms.accessRights | open access | en_US |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/14535