Pub Date : 2022-05-16DOI: 10.23919/ondm54585.2022.9782815
Saquib Amjad, S. Patri, C. M. Machuca
In optical networks, the reach of the optical signal is controlled by the receiver’s capability to successfully receive the signal, degraded due to optical impairments and noise. This reach can be extended by using regeneration at intermediate nodes. Efficient placement and minimization of the number of regenerators is referred to as the regenerator placement problem. This paper proposes a method to solve the regenerator placement problem in a multiperiod planning scenario with the objective of maximizing throughput with minimum lightpaths. The paper addresses regenerator placement in two phases, a preselection of possible locations for regeneration based on OSNR constraints, and provisioning a combination of regenerated and non-regenerated lightpaths. The provisioning formulation focuses on minimizing the number of transceivers while maximizing the datarate. We demonstrate the advantage of our approach compared to state-of-the-art methods in terms of throughput, underprovisioning and number of transceivers on 3 different topologies. Our results show that the proposed solution is able to meet the dynamic traffic with lower underprovisioning.
{"title":"Towards Regeneration in Flexible Optical Network Planning","authors":"Saquib Amjad, S. Patri, C. M. Machuca","doi":"10.23919/ondm54585.2022.9782815","DOIUrl":"https://doi.org/10.23919/ondm54585.2022.9782815","url":null,"abstract":"In optical networks, the reach of the optical signal is controlled by the receiver’s capability to successfully receive the signal, degraded due to optical impairments and noise. This reach can be extended by using regeneration at intermediate nodes. Efficient placement and minimization of the number of regenerators is referred to as the regenerator placement problem. This paper proposes a method to solve the regenerator placement problem in a multiperiod planning scenario with the objective of maximizing throughput with minimum lightpaths. The paper addresses regenerator placement in two phases, a preselection of possible locations for regeneration based on OSNR constraints, and provisioning a combination of regenerated and non-regenerated lightpaths. The provisioning formulation focuses on minimizing the number of transceivers while maximizing the datarate. We demonstrate the advantage of our approach compared to state-of-the-art methods in terms of throughput, underprovisioning and number of transceivers on 3 different topologies. Our results show that the proposed solution is able to meet the dynamic traffic with lower underprovisioning.","PeriodicalId":191683,"journal":{"name":"2022 International Conference on Optical Network Design and Modeling (ONDM)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117139486","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-16DOI: 10.23919/ondm54585.2022.9782848
P. Boffi, M. Ferrario, I. D. Luch, G. Rizzelli, R. Gaudino
The telecommunication fiber network already deployed in urban areas provides an added value to the optical asset itself, allowing a smart monitoring of our cities in a large scale. It is possible to use deployed PON infrastructures for structural vibration and local seismologic perturbations monitoring. On the other hand, surveillance of the embedded network and real-time safety diagnostic is also possible. The invited talk will present different experimental demonstrations to show the sensing performance by exploiting deployed fiber links, assessing the compatibility with the optical data telecom traffic at very high rate.
{"title":"Optical sensing in urban areas by deployed telecommunication fiber networks","authors":"P. Boffi, M. Ferrario, I. D. Luch, G. Rizzelli, R. Gaudino","doi":"10.23919/ondm54585.2022.9782848","DOIUrl":"https://doi.org/10.23919/ondm54585.2022.9782848","url":null,"abstract":"The telecommunication fiber network already deployed in urban areas provides an added value to the optical asset itself, allowing a smart monitoring of our cities in a large scale. It is possible to use deployed PON infrastructures for structural vibration and local seismologic perturbations monitoring. On the other hand, surveillance of the embedded network and real-time safety diagnostic is also possible. The invited talk will present different experimental demonstrations to show the sensing performance by exploiting deployed fiber links, assessing the compatibility with the optical data telecom traffic at very high rate.","PeriodicalId":191683,"journal":{"name":"2022 International Conference on Optical Network Design and Modeling (ONDM)","volume":"C-24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126477823","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-16DOI: 10.23919/ondm54585.2022.9782838
S. Singh, Che-Yu Liu, S. Yoo, R. Proietti
The high bandwidth and low latency requirements of modern computing applications with their dynamic and nonuniform traffic patterns impose severe challenges to current data center (DC) and high performance computing (HPC) networks. Therefore, we present a dynamic network reconfiguration mechanism that could satisfy the time-varying applications’ demands in an optical DC/HPC network. We propose a direct and an indirect topology extraction methods based on a machine learning-aided traffic prediction approach under multi-application scenario. The traffic prediction for topology extraction and bandwidth reconfiguration (PredicTER) method could lead to frequent topology and bandwidth reconfiguration. In contrast, the indirect approach, namely traffic prediction with clustering for topology extraction and bandwidth reconfiguration (PrediCLUSTER), utilizes an unsupervised learning-based clustering model to first associate the predicted traffic to one of possible traffic clusters, and then extracts a common topology for the cluster. This restricts the reconfigured topology set to the number of traffic clusters. Our simulation results show that the time-average of mean packet latencies (and total dropped packets) over 60 seconds of timevarying traffic under the PredicTER, PrediCLUSTER and a static topology are 37.7μs,41.2μs, and 50.2μs (and 37,967, 12,305, and 36,836), respectively. Overall, the PredicTER (and PrediCLUSTER) method(s) can improve the end-to-end packet latency by 24.9% (and 17.8%), and the packet loss rate by −3.1% (and 66.6%), as compared to the static flat Hyper-X-like topology.
{"title":"Machine-Learning-Aided Dynamic Reconfiguration in Optical DC/HPC Networks (Invited)","authors":"S. Singh, Che-Yu Liu, S. Yoo, R. Proietti","doi":"10.23919/ondm54585.2022.9782838","DOIUrl":"https://doi.org/10.23919/ondm54585.2022.9782838","url":null,"abstract":"The high bandwidth and low latency requirements of modern computing applications with their dynamic and nonuniform traffic patterns impose severe challenges to current data center (DC) and high performance computing (HPC) networks. Therefore, we present a dynamic network reconfiguration mechanism that could satisfy the time-varying applications’ demands in an optical DC/HPC network. We propose a direct and an indirect topology extraction methods based on a machine learning-aided traffic prediction approach under multi-application scenario. The traffic prediction for topology extraction and bandwidth reconfiguration (PredicTER) method could lead to frequent topology and bandwidth reconfiguration. In contrast, the indirect approach, namely traffic prediction with clustering for topology extraction and bandwidth reconfiguration (PrediCLUSTER), utilizes an unsupervised learning-based clustering model to first associate the predicted traffic to one of possible traffic clusters, and then extracts a common topology for the cluster. This restricts the reconfigured topology set to the number of traffic clusters. Our simulation results show that the time-average of mean packet latencies (and total dropped packets) over 60 seconds of timevarying traffic under the PredicTER, PrediCLUSTER and a static topology are 37.7μs,41.2μs, and 50.2μs (and 37,967, 12,305, and 36,836), respectively. Overall, the PredicTER (and PrediCLUSTER) method(s) can improve the end-to-end packet latency by 24.9% (and 17.8%), and the packet loss rate by −3.1% (and 66.6%), as compared to the static flat Hyper-X-like topology.","PeriodicalId":191683,"journal":{"name":"2022 International Conference on Optical Network Design and Modeling (ONDM)","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125482697","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-16DOI: 10.23919/ondm54585.2022.9782851
J. Pesic, Marina Curtol, Lahcen Abnaou, Abdelali El Imadi, Stefano Morganti
This paper provides an overview of the missing pieces currently preventing effective application of machine learning in the field. We discuss access to field data and we perform a proof of concept for the two SDN automation use cases based on programmable hardware, open APIs and streaming telemetry. The automation workflow with its performance evaluations is also presented
{"title":"SDN Automation for Optical Networks Based on Open APIs and Streaming Telemetry","authors":"J. Pesic, Marina Curtol, Lahcen Abnaou, Abdelali El Imadi, Stefano Morganti","doi":"10.23919/ondm54585.2022.9782851","DOIUrl":"https://doi.org/10.23919/ondm54585.2022.9782851","url":null,"abstract":"This paper provides an overview of the missing pieces currently preventing effective application of machine learning in the field. We discuss access to field data and we perform a proof of concept for the two SDN automation use cases based on programmable hardware, open APIs and streaming telemetry. The automation workflow with its performance evaluations is also presented","PeriodicalId":191683,"journal":{"name":"2022 International Conference on Optical Network Design and Modeling (ONDM)","volume":"259 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121443483","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-16DOI: 10.23919/ondm54585.2022.9782839
Srivatsan Balasubramanian, B. Gangopadhyay, V. Dangui, S. Ahuja, Varun Gupta, G. Pastukhov, Max Noormohammadpour, A. Nikolaidis, Ariyani Copley, Xueqi He, Jiachuan Tian, Jiajia Chen, Arash Vakili, Chiunlin Lim, Guanqing Yan, Anand Gokul, Biao Lu, Debottym Mukherjee
Meta has a large scale backbone infrastructure supporting services with varying QoS requirements. As part of backbone network planning, a capacity plan that differentiates between different classes of services in terms of availability guarantees is generated and scheduled for deployment. Deployment progress is measured traditionally in terms of volumes of capacity deployed. Our work provides insights into the shortcomings of capacity volume driven deployments. We provide a methodology to rank the contribution of each entity pending deployment towards our network performance goals and use this metric to prioritize deployments helping higher classes of services meet their network guarantees earlier in the deployment schedule. By enabling QoS awareness in backbone deployments, we are able to demonstrate a 67% reduction of risk exposure period for high priority services.
{"title":"Prioritizing deployments achieving targeted network performance across a multilayer Pb/s network","authors":"Srivatsan Balasubramanian, B. Gangopadhyay, V. Dangui, S. Ahuja, Varun Gupta, G. Pastukhov, Max Noormohammadpour, A. Nikolaidis, Ariyani Copley, Xueqi He, Jiachuan Tian, Jiajia Chen, Arash Vakili, Chiunlin Lim, Guanqing Yan, Anand Gokul, Biao Lu, Debottym Mukherjee","doi":"10.23919/ondm54585.2022.9782839","DOIUrl":"https://doi.org/10.23919/ondm54585.2022.9782839","url":null,"abstract":"Meta has a large scale backbone infrastructure supporting services with varying QoS requirements. As part of backbone network planning, a capacity plan that differentiates between different classes of services in terms of availability guarantees is generated and scheduled for deployment. Deployment progress is measured traditionally in terms of volumes of capacity deployed. Our work provides insights into the shortcomings of capacity volume driven deployments. We provide a methodology to rank the contribution of each entity pending deployment towards our network performance goals and use this metric to prioritize deployments helping higher classes of services meet their network guarantees earlier in the deployment schedule. By enabling QoS awareness in backbone deployments, we are able to demonstrate a 67% reduction of risk exposure period for high priority services.","PeriodicalId":191683,"journal":{"name":"2022 International Conference on Optical Network Design and Modeling (ONDM)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122075513","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-16DOI: 10.23919/ondm54585.2022.9782866
M. Iqbal, Luis Velasco, M. Ruiz, A. Napoli, J. Pedro, N. Costa
Quantum internet, which is expected to be a combination of quantum and classical networks, promises to provide information-theoretic security for data exchange. Classical networks have well-established protocols for reliable end-to-end transmission that implicitly make use of duplicating classical bits. However, quantum bits (qubits) cannot be copied due to the no-cloning theorem. In this paper, we take advantage of the principle of creating imperfect clones using a Universal Quantum Copying Machine (UQCM) and propose the Quantum Automatic Repeat Request (QARQ) protocol, inspired by its classical equivalent. A simulation platform has been developed to study the feasibility of QARQ. Results show that our proposal is well suited for applications that are compatible with low fidelity requirements.
{"title":"Quantum Bit Retransmission Using Universal Quantum Copying Machine","authors":"M. Iqbal, Luis Velasco, M. Ruiz, A. Napoli, J. Pedro, N. Costa","doi":"10.23919/ondm54585.2022.9782866","DOIUrl":"https://doi.org/10.23919/ondm54585.2022.9782866","url":null,"abstract":"Quantum internet, which is expected to be a combination of quantum and classical networks, promises to provide information-theoretic security for data exchange. Classical networks have well-established protocols for reliable end-to-end transmission that implicitly make use of duplicating classical bits. However, quantum bits (qubits) cannot be copied due to the no-cloning theorem. In this paper, we take advantage of the principle of creating imperfect clones using a Universal Quantum Copying Machine (UQCM) and propose the Quantum Automatic Repeat Request (QARQ) protocol, inspired by its classical equivalent. A simulation platform has been developed to study the feasibility of QARQ. Results show that our proposal is well suited for applications that are compatible with low fidelity requirements.","PeriodicalId":191683,"journal":{"name":"2022 International Conference on Optical Network Design and Modeling (ONDM)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129401973","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-16DOI: 10.23919/ondm54585.2022.9782856
Fehmida Usmani, I. Khan, M. U. Masood, Arsalan Ahmad, Muhammad Shahzad, V. Curri
The current increase in bandwidth-hungry applications and the progressively evolving concept of connected "smart" devices through the internet have increased internet traffic exponentially. To hold this expansion of internet traffic, the network operators insist on the full capacity utilization of already deployed hardware infrastructure. In this context, accurate and earlier calculation of the quality of transmission (QoT) of the lightpaths (LPs) is critical for minimizing the required margins that arise due to the uncertainty in the operating point of network elements. This article proposes a novel framework in which a transfer learning assisted QoT-Estimation (QoT-E) is made. The transfer learning agent acquired the knowledge from a traditional fully operational network operating on C-band and utilized this knowledge to assist the operator in estimating the LP QoT on a state-of-the-art newly functioning network on an extended C-band operating with 400ZR standards. The measurement parameter considered to estimate the QoT of LP is the generalized signal-to-noise ratio (GSNR). The dataset used in this analysis is generated synthetically by utilizing well tested GNPy platform. Promising results are achieved in terms of reducing the overall required margin and better utilization of the residual network capacity.
{"title":"Transfer learning Aided QoT Computation in Network Operating with the 400ZR Standard","authors":"Fehmida Usmani, I. Khan, M. U. Masood, Arsalan Ahmad, Muhammad Shahzad, V. Curri","doi":"10.23919/ondm54585.2022.9782856","DOIUrl":"https://doi.org/10.23919/ondm54585.2022.9782856","url":null,"abstract":"The current increase in bandwidth-hungry applications and the progressively evolving concept of connected \"smart\" devices through the internet have increased internet traffic exponentially. To hold this expansion of internet traffic, the network operators insist on the full capacity utilization of already deployed hardware infrastructure. In this context, accurate and earlier calculation of the quality of transmission (QoT) of the lightpaths (LPs) is critical for minimizing the required margins that arise due to the uncertainty in the operating point of network elements. This article proposes a novel framework in which a transfer learning assisted QoT-Estimation (QoT-E) is made. The transfer learning agent acquired the knowledge from a traditional fully operational network operating on C-band and utilized this knowledge to assist the operator in estimating the LP QoT on a state-of-the-art newly functioning network on an extended C-band operating with 400ZR standards. The measurement parameter considered to estimate the QoT of LP is the generalized signal-to-noise ratio (GSNR). The dataset used in this analysis is generated synthetically by utilizing well tested GNPy platform. Promising results are achieved in terms of reducing the overall required margin and better utilization of the residual network capacity.","PeriodicalId":191683,"journal":{"name":"2022 International Conference on Optical Network Design and Modeling (ONDM)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128915368","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
For many years the best strategy to optimize the Total Cost of Ownership (TCO) of an IP+Optical network has been to reduce as much as possible the utilization of IP routers’ switch fabrics and interfaces. This can be achieved by means of optical bypass using Reconfigurable Add/Drop Multiplexers (ROADMs). This strategy comes at the cost of a suboptimal wavelength utilization and longer (on average) optical links, running with a lower OSNR. In this paper we analyse alternative architectures which take advantage of the latest Network Processing Units (NPUs) in IP routers and pluggable 400G DWDM interfaces, which helps reducing the cost associated to packet processing.
{"title":"Routed Optical Networking: an alternative architecture for IP+Optical aggregation networks","authors":"Valerio Viscardi, Dirk Schroetter, Moustafa Kattan","doi":"10.23919/ondm54585.2022.9782858","DOIUrl":"https://doi.org/10.23919/ondm54585.2022.9782858","url":null,"abstract":"For many years the best strategy to optimize the Total Cost of Ownership (TCO) of an IP+Optical network has been to reduce as much as possible the utilization of IP routers’ switch fabrics and interfaces. This can be achieved by means of optical bypass using Reconfigurable Add/Drop Multiplexers (ROADMs). This strategy comes at the cost of a suboptimal wavelength utilization and longer (on average) optical links, running with a lower OSNR. In this paper we analyse alternative architectures which take advantage of the latest Network Processing Units (NPUs) in IP routers and pluggable 400G DWDM interfaces, which helps reducing the cost associated to packet processing.","PeriodicalId":191683,"journal":{"name":"2022 International Conference on Optical Network Design and Modeling (ONDM)","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114076741","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-16DOI: 10.23919/ondm54585.2022.9782845
T. Miyamura, A. Misawa
We propose an adaptive joint optimization method of IT resources and optical spectrum under time-varying traffic demand in elastic optical networks while avoiding an increase in operation cost. Currently, numerous network services are provided by a service function chain (SFC). Once SFCs are provisioned, an optical path is established to connect the SFC and users. SFCs are placed in one of the candidate datacenters in the network by considering residual IT resources and the location of users. Here, the optimal placement of SFCs can vary due to service demand changes. To maintain network performance, we need to reconfigure network configuration by migrating SFCs and rerouting optical paths. However, such reconfiguration requires additional operation cost. In this paper, we consider the joint optimization problem of IT resources and optical spectrum in consideration of operation cost. We formulate the problem as mixed integer linear programming and then quantitatively evaluate the trade-off relationship between the optimality of reconfiguration and operation cost. We demonstrate that we can achieve sufficient network performance through the adaptive joint optimization while suppressing an increase in operation cost.
{"title":"Adaptive Joint Optimization of IT Resources and Optical Spectrum Considering Operation Cost","authors":"T. Miyamura, A. Misawa","doi":"10.23919/ondm54585.2022.9782845","DOIUrl":"https://doi.org/10.23919/ondm54585.2022.9782845","url":null,"abstract":"We propose an adaptive joint optimization method of IT resources and optical spectrum under time-varying traffic demand in elastic optical networks while avoiding an increase in operation cost. Currently, numerous network services are provided by a service function chain (SFC). Once SFCs are provisioned, an optical path is established to connect the SFC and users. SFCs are placed in one of the candidate datacenters in the network by considering residual IT resources and the location of users. Here, the optimal placement of SFCs can vary due to service demand changes. To maintain network performance, we need to reconfigure network configuration by migrating SFCs and rerouting optical paths. However, such reconfiguration requires additional operation cost. In this paper, we consider the joint optimization problem of IT resources and optical spectrum in consideration of operation cost. We formulate the problem as mixed integer linear programming and then quantitatively evaluate the trade-off relationship between the optimality of reconfiguration and operation cost. We demonstrate that we can achieve sufficient network performance through the adaptive joint optimization while suppressing an increase in operation cost.","PeriodicalId":191683,"journal":{"name":"2022 International Conference on Optical Network Design and Modeling (ONDM)","volume":"315 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116531178","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-16DOI: 10.23919/ondm54585.2022.9782850
Aleksandra Knapińska, K. Póltorak, Dominika Poreba, Jan Miszczyk, Mateusz Daniluk, K. Walkowiak
The knowledge about future traffic volumes is beneficial for the network operators in many areas. Short-term forecasting of multiple traffic types helps with efficient resource utilization by enabling near real-time adjustment. An important issue is the choice of a suitable prediction model to obtain the most accurate traffic forecasts. A machine learning (ML) algorithm picked for this task can be further tuned by an appropriate feature selection. In this paper, we propose three models containing sets of additional input features to improve the prediction quality of different ML algorithms. We evaluate our models on multiple datasets containing diverse types of network traffic. In extensive numerical experiments, we prove the high prediction quality of ML regression algorithms aided by our proposed additional features. Obtained mean absolute percentage errors (MAPE) are, depending on the predicted traffic type, as little as 1–10%.
{"title":"On Feature Selection in Short-Term Prediction of Backbone Optical Network Traffic","authors":"Aleksandra Knapińska, K. Póltorak, Dominika Poreba, Jan Miszczyk, Mateusz Daniluk, K. Walkowiak","doi":"10.23919/ondm54585.2022.9782850","DOIUrl":"https://doi.org/10.23919/ondm54585.2022.9782850","url":null,"abstract":"The knowledge about future traffic volumes is beneficial for the network operators in many areas. Short-term forecasting of multiple traffic types helps with efficient resource utilization by enabling near real-time adjustment. An important issue is the choice of a suitable prediction model to obtain the most accurate traffic forecasts. A machine learning (ML) algorithm picked for this task can be further tuned by an appropriate feature selection. In this paper, we propose three models containing sets of additional input features to improve the prediction quality of different ML algorithms. We evaluate our models on multiple datasets containing diverse types of network traffic. In extensive numerical experiments, we prove the high prediction quality of ML regression algorithms aided by our proposed additional features. Obtained mean absolute percentage errors (MAPE) are, depending on the predicted traffic type, as little as 1–10%.","PeriodicalId":191683,"journal":{"name":"2022 International Conference on Optical Network Design and Modeling (ONDM)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115399940","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：