Pub Date : 2023-06-19DOI: 10.1109/NetSoft57336.2023.10175449
Jacopo Massa, Stefano Forti, F. Paganelli, Patrizio Dazzi, Antonio Brogi
Intent-based Networking (IBN) aims at simplifying network configuration and management by using high-level objectives that express the desired state of the network rather than the details of how to implement it. In this article, we propose a declarative methodology and an associated open-source Prolog prototype (i) to model IBN intents related to the provisioning of Virtual Network Function (VNF) chains, and (ii) to process those intents to assemble and place a VNF chain that fulfils them. Our prototype is assessed over a lifelike motivating scenario.
{"title":"Declarative Provisioning of Virtual Network Function Chains in Intent-based Networks","authors":"Jacopo Massa, Stefano Forti, F. Paganelli, Patrizio Dazzi, Antonio Brogi","doi":"10.1109/NetSoft57336.2023.10175449","DOIUrl":"https://doi.org/10.1109/NetSoft57336.2023.10175449","url":null,"abstract":"Intent-based Networking (IBN) aims at simplifying network configuration and management by using high-level objectives that express the desired state of the network rather than the details of how to implement it. In this article, we propose a declarative methodology and an associated open-source Prolog prototype (i) to model IBN intents related to the provisioning of Virtual Network Function (VNF) chains, and (ii) to process those intents to assemble and place a VNF chain that fulfils them. Our prototype is assessed over a lifelike motivating scenario.","PeriodicalId":223208,"journal":{"name":"2023 IEEE 9th International Conference on Network Softwarization (NetSoft)","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123315551","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 : 2023-06-19DOI: 10.1109/NetSoft57336.2023.10175464
Aristide T.-J. Akem, Beyza Bütün, Michele Gucciardo, M. Fiore
Recent endeavours have enabled the integration of trained machine learning models like Random Forests in resource-constrained programmable switches for line rate inference. In this work, we first show how packet-level information can be used to classify individual packets in production-level hardware with very low latency. We then demonstrate how the newly proposed Flowrest framework improves classification performance relative to the packet-level approach by exploiting flow-level statistics to instead classify traffic flows entirely within the switch without considerably increasing latency. We conduct experiments using measurement data in a real-world testbed with an Intel Tofino switch and shed light on how Flowrest achieves an F1-score of 99% in a service classification use case, outperforming its packet-level counterpart by 8%.
{"title":"Showcasing In-Switch Machine Learning Inference","authors":"Aristide T.-J. Akem, Beyza Bütün, Michele Gucciardo, M. Fiore","doi":"10.1109/NetSoft57336.2023.10175464","DOIUrl":"https://doi.org/10.1109/NetSoft57336.2023.10175464","url":null,"abstract":"Recent endeavours have enabled the integration of trained machine learning models like Random Forests in resource-constrained programmable switches for line rate inference. In this work, we first show how packet-level information can be used to classify individual packets in production-level hardware with very low latency. We then demonstrate how the newly proposed Flowrest framework improves classification performance relative to the packet-level approach by exploiting flow-level statistics to instead classify traffic flows entirely within the switch without considerably increasing latency. We conduct experiments using measurement data in a real-world testbed with an Intel Tofino switch and shed light on how Flowrest achieves an F1-score of 99% in a service classification use case, outperforming its packet-level counterpart by 8%.","PeriodicalId":223208,"journal":{"name":"2023 IEEE 9th International Conference on Network Softwarization (NetSoft)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114528217","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 : 2023-06-19DOI: 10.1109/NetSoft57336.2023.10175483
H. Shimonishi, M. Murata, G. Hasegawa, Nattaon Techasarntikul
For the future cyber-physical system (CPS) society, it is necessary to construct digital twins (DTs) of a real world in real time using a lot of cameras and sensors. Hence, the energy efficiency of both networks and computers for largescale distributed video analysis is a major challenge for the full-scale spread of CPSs and DTs. Toward this goal, we first propose a model to arbitrarily split and distribute the video analysis task to terminals, edge servers, and cloud servers and dynamically assign appropriate CNN models to them. System-wide optimization of such distributed processing can reduce overall system power consumption by reducing network bandwidth and efficiently utilizing distributed CPU/GPU resources. To realize this optimization in a real system, we also propose a model to estimate the GPU load, processing time, and power consumption of these devices based on massive experimental measurements. Since such a large-scale optimization is difficult because of the dynamic and multi-objective nature of the problem, we propose a new optimization algorithm composed of Genetic Algorithm and Bayesian Attractor Model. Finally, simulation evaluations are performed to demonstrate that the proposed method can minimize system power consumption and satisfy latency and recognition accuracy requirements of each video analysis, even under changing environmental conditions.
{"title":"Energy Optimization of Distributed Video Processing System using Genetic Algorithm with Bayesian Attractor Model","authors":"H. Shimonishi, M. Murata, G. Hasegawa, Nattaon Techasarntikul","doi":"10.1109/NetSoft57336.2023.10175483","DOIUrl":"https://doi.org/10.1109/NetSoft57336.2023.10175483","url":null,"abstract":"For the future cyber-physical system (CPS) society, it is necessary to construct digital twins (DTs) of a real world in real time using a lot of cameras and sensors. Hence, the energy efficiency of both networks and computers for largescale distributed video analysis is a major challenge for the full-scale spread of CPSs and DTs. Toward this goal, we first propose a model to arbitrarily split and distribute the video analysis task to terminals, edge servers, and cloud servers and dynamically assign appropriate CNN models to them. System-wide optimization of such distributed processing can reduce overall system power consumption by reducing network bandwidth and efficiently utilizing distributed CPU/GPU resources. To realize this optimization in a real system, we also propose a model to estimate the GPU load, processing time, and power consumption of these devices based on massive experimental measurements. Since such a large-scale optimization is difficult because of the dynamic and multi-objective nature of the problem, we propose a new optimization algorithm composed of Genetic Algorithm and Bayesian Attractor Model. Finally, simulation evaluations are performed to demonstrate that the proposed method can minimize system power consumption and satisfy latency and recognition accuracy requirements of each video analysis, even under changing environmental conditions.","PeriodicalId":223208,"journal":{"name":"2023 IEEE 9th International Conference on Network Softwarization (NetSoft)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131764356","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 : 2023-06-19DOI: 10.1109/NetSoft57336.2023.10175484
Eliana Neuza Silva, Fernando Mira da Silva
Smart vehicles in Vehicular Edge Computing Environments run latency sensitive applications, such as driver assistance, autonomous driving, accident prevention and others that require quick response times due to low latency constraints. This work focus on the workload orchestration and the decision to offload vehicular application tasks from vehicles to the network edge to increase computing powers and minimize latency. We introduce a new offloading orchestration algorithm based on Deep Reinforcement Learning. We show that the proposed algorithm has a lower task failure rate than the best solutions from the literature, while requiring lower computational power.
{"title":"Deep Reinforcement Learning Edge Workload Orchestrator for Vehicular Edge Computing","authors":"Eliana Neuza Silva, Fernando Mira da Silva","doi":"10.1109/NetSoft57336.2023.10175484","DOIUrl":"https://doi.org/10.1109/NetSoft57336.2023.10175484","url":null,"abstract":"Smart vehicles in Vehicular Edge Computing Environments run latency sensitive applications, such as driver assistance, autonomous driving, accident prevention and others that require quick response times due to low latency constraints. This work focus on the workload orchestration and the decision to offload vehicular application tasks from vehicles to the network edge to increase computing powers and minimize latency. We introduce a new offloading orchestration algorithm based on Deep Reinforcement Learning. We show that the proposed algorithm has a lower task failure rate than the best solutions from the literature, while requiring lower computational power.","PeriodicalId":223208,"journal":{"name":"2023 IEEE 9th International Conference on Network Softwarization (NetSoft)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116992801","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 : 2023-06-19DOI: 10.1109/NetSoft57336.2023.10175458
Annalisa Navarro, R. Canonico, A. Botta
Software Defined Wide Area Network (SD-WAN) is rapidly becoming an attractive solution for enterprise networks as it offers several benefits such as cost efficiency, increased bandwidth, and improved application performance. However, SDWAN also brings new challenges that must be addressed for effective deployment (i.e. openness, interoperability, network automation, monitoring, QoS guarantees, scalability and security). In this paper, we highlight the criticalities of this technology and analyze the solutions proposed by the state of the art. We then present a scalable framework based on distributed Reinforcement Learning agents for guaranteeing availability and QoS to business applications. We believe that our work provides valuable insights into the opportunities and challenges of SD-WAN technology and offers new perspectives for future research in this area.
{"title":"Software Defined Wide Area Networks: Current Challenges and Future Perspectives","authors":"Annalisa Navarro, R. Canonico, A. Botta","doi":"10.1109/NetSoft57336.2023.10175458","DOIUrl":"https://doi.org/10.1109/NetSoft57336.2023.10175458","url":null,"abstract":"Software Defined Wide Area Network (SD-WAN) is rapidly becoming an attractive solution for enterprise networks as it offers several benefits such as cost efficiency, increased bandwidth, and improved application performance. However, SDWAN also brings new challenges that must be addressed for effective deployment (i.e. openness, interoperability, network automation, monitoring, QoS guarantees, scalability and security). In this paper, we highlight the criticalities of this technology and analyze the solutions proposed by the state of the art. We then present a scalable framework based on distributed Reinforcement Learning agents for guaranteeing availability and QoS to business applications. We believe that our work provides valuable insights into the opportunities and challenges of SD-WAN technology and offers new perspectives for future research in this area.","PeriodicalId":223208,"journal":{"name":"2023 IEEE 9th International Conference on Network Softwarization (NetSoft)","volume":"54 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123940947","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 : 2023-06-19DOI: 10.1109/NetSoft57336.2023.10175403
Paulo Araújo, H. López, João Faria, Alexandre J. T. Santos
Multi-access Edge Computing (MEC) brings cloud capabilities to the edge, allowing less powerful devices to offload tasks while, together with the effective distribution of content and resources across edge hosts, enabling ultra-low latency. With the distributed resources across the network and mobile devices constantly changing location, it is most important to correctly manage and orchestrate those resources to deliver the best possible network performance efficiently. This paper investigates how MEC systems manage mobility and which efforts have been made in mobility simulation in MEC. A MEC simulation setup and two scenarios were developed using Simu5G and evaluated regarding their latency and use of resources when subject to SUMO-generated mobility data. Analysis of the simulated scenarios show that the MEC systems reacts differently depending upon different mobility conditions. Also, the developed setup showed to be very useful to develop and test mobility management strategies in MEC environments.
{"title":"Towards Mobility Management in MEC Simulation","authors":"Paulo Araújo, H. López, João Faria, Alexandre J. T. Santos","doi":"10.1109/NetSoft57336.2023.10175403","DOIUrl":"https://doi.org/10.1109/NetSoft57336.2023.10175403","url":null,"abstract":"Multi-access Edge Computing (MEC) brings cloud capabilities to the edge, allowing less powerful devices to offload tasks while, together with the effective distribution of content and resources across edge hosts, enabling ultra-low latency. With the distributed resources across the network and mobile devices constantly changing location, it is most important to correctly manage and orchestrate those resources to deliver the best possible network performance efficiently. This paper investigates how MEC systems manage mobility and which efforts have been made in mobility simulation in MEC. A MEC simulation setup and two scenarios were developed using Simu5G and evaluated regarding their latency and use of resources when subject to SUMO-generated mobility data. Analysis of the simulated scenarios show that the MEC systems reacts differently depending upon different mobility conditions. Also, the developed setup showed to be very useful to develop and test mobility management strategies in MEC environments.","PeriodicalId":223208,"journal":{"name":"2023 IEEE 9th International Conference on Network Softwarization (NetSoft)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128367821","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 : 2023-06-19DOI: 10.1109/NetSoft57336.2023.10175447
Angel M. Gama Garcia, J. A. Calero, H. Mora, Q. Wang
With the evolution of softwarised industrial infrastructures, there is an increasing need for more sophisticated cyber security solutions that can protect industrial processes from a rapidly evolving landscape of cyber threats. In this context, we present an agent-based approach that provides process monitoring, predictive process behaviour, and process control to give the organisations appropriate situational awareness in relation to cyber security threats, enabling them to re-actively or pro-actively detect attacks and respond to advanced persistent threats and multi-vector attacks. Our architectural solution is based on four agents: Process Inventory Agent (PIA), Process Monitoring Agent (PMA), Process Forecasting Agent (PFA), and the Process Slicing Control Agent (PSCA), which work together to deliver a novel mitigation tool to secure softwarised industrial environments. The architecture has been designed, prototyped, and validated in order to demonstrate the effectiveness of our solution. Experimental results show that the proposed solution can successfully mitigate different attacks in the concerned context.
{"title":"Process Slicing: A New Mitigation Tool for Cyber-attacks against Softwarised Industrial Environments","authors":"Angel M. Gama Garcia, J. A. Calero, H. Mora, Q. Wang","doi":"10.1109/NetSoft57336.2023.10175447","DOIUrl":"https://doi.org/10.1109/NetSoft57336.2023.10175447","url":null,"abstract":"With the evolution of softwarised industrial infrastructures, there is an increasing need for more sophisticated cyber security solutions that can protect industrial processes from a rapidly evolving landscape of cyber threats. In this context, we present an agent-based approach that provides process monitoring, predictive process behaviour, and process control to give the organisations appropriate situational awareness in relation to cyber security threats, enabling them to re-actively or pro-actively detect attacks and respond to advanced persistent threats and multi-vector attacks. Our architectural solution is based on four agents: Process Inventory Agent (PIA), Process Monitoring Agent (PMA), Process Forecasting Agent (PFA), and the Process Slicing Control Agent (PSCA), which work together to deliver a novel mitigation tool to secure softwarised industrial environments. The architecture has been designed, prototyped, and validated in order to demonstrate the effectiveness of our solution. Experimental results show that the proposed solution can successfully mitigate different attacks in the concerned context.","PeriodicalId":223208,"journal":{"name":"2023 IEEE 9th International Conference on Network Softwarization (NetSoft)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128598421","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 : 2023-06-19DOI: 10.1109/NetSoft57336.2023.10175463
F. Vogt, F. R. Cesen, Ariel Góes deCastro, M. C. Luizelli, Christian Esteve Rothenberg, Gergely Pongrácz
In recent years, advances in virtual reality (VR) technologies (e.g., high-quality VR headsets) have enabled a new perspective of experiences for users (e.g., gaming, online events). However, ensuring the user experience is still a challenge. Existing solutions are limited to measuring and estimating QoE at the user plane (e.g., VR player) or at the control plane, imposing unfeasible latency for different scenarios (5G networks and beyond). In this work, we propose QoEyes, an in-network QoE estimation based on the use of Inter-Packet-Gap (IPG) in programmable devices. Our results show that the IPG measured on the data plane is strongly linked to QoE, yielding an accurate data plane QoE estimate.
{"title":"QoEyes: Towards Virtual Reality Streaming QoE Estimation Entirely in the Data Plane","authors":"F. Vogt, F. R. Cesen, Ariel Góes deCastro, M. C. Luizelli, Christian Esteve Rothenberg, Gergely Pongrácz","doi":"10.1109/NetSoft57336.2023.10175463","DOIUrl":"https://doi.org/10.1109/NetSoft57336.2023.10175463","url":null,"abstract":"In recent years, advances in virtual reality (VR) technologies (e.g., high-quality VR headsets) have enabled a new perspective of experiences for users (e.g., gaming, online events). However, ensuring the user experience is still a challenge. Existing solutions are limited to measuring and estimating QoE at the user plane (e.g., VR player) or at the control plane, imposing unfeasible latency for different scenarios (5G networks and beyond). In this work, we propose QoEyes, an in-network QoE estimation based on the use of Inter-Packet-Gap (IPG) in programmable devices. Our results show that the IPG measured on the data plane is strongly linked to QoE, yielding an accurate data plane QoE estimate.","PeriodicalId":223208,"journal":{"name":"2023 IEEE 9th International Conference on Network Softwarization (NetSoft)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127155349","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 : 2023-06-19DOI: 10.1109/NetSoft57336.2023.10175492
Prateek Bagora, Amin Ebrahimzadeh, F. Wuhib, R. Glitho
Ensuring the quality of service of applications deployed in inherently complex and fault-prone cloud environments is of utmost concern. While machine learning based fault management solutions help attain the desired reliability, they require labeled cloud metrics data for training and evaluation. Furthermore, high dynamicity of cloud environments brings forth emerging data distributions, which necessitate frequent labeling of data for model adaptation. We propose a test suite-based active learning framework for automated labeling of cloud metrics data with the corresponding cloud system state while accounting for emerging fault patterns and data or concept drifts. We have implemented our solution on a cloud testbed and introduced various emerging data distribution scenarios to evaluate the proposed framework’s labeling efficacy over known and emerging data distributions. According to our results, the proposed framework achieves a 41% higher weighted Fl-score and a 34% higher average AUC score than a system without any adaptation for emerging data distributions.
{"title":"Data Labeling for Fault Detection in Cloud: A Test Suite-Based Active Learning Approach","authors":"Prateek Bagora, Amin Ebrahimzadeh, F. Wuhib, R. Glitho","doi":"10.1109/NetSoft57336.2023.10175492","DOIUrl":"https://doi.org/10.1109/NetSoft57336.2023.10175492","url":null,"abstract":"Ensuring the quality of service of applications deployed in inherently complex and fault-prone cloud environments is of utmost concern. While machine learning based fault management solutions help attain the desired reliability, they require labeled cloud metrics data for training and evaluation. Furthermore, high dynamicity of cloud environments brings forth emerging data distributions, which necessitate frequent labeling of data for model adaptation. We propose a test suite-based active learning framework for automated labeling of cloud metrics data with the corresponding cloud system state while accounting for emerging fault patterns and data or concept drifts. We have implemented our solution on a cloud testbed and introduced various emerging data distribution scenarios to evaluate the proposed framework’s labeling efficacy over known and emerging data distributions. According to our results, the proposed framework achieves a 41% higher weighted Fl-score and a 34% higher average AUC score than a system without any adaptation for emerging data distributions.","PeriodicalId":223208,"journal":{"name":"2023 IEEE 9th International Conference on Network Softwarization (NetSoft)","volume":"236 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130089215","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 : 2023-06-19DOI: 10.1109/NetSoft57336.2023.10175441
Marcos Carvalho, Daniel Soares, D. Macedo
Cloud Gaming renders game data in the cloud and forwards it to players over the network. While this reduces hardware costs for players, it introduces challenges in network management and delivering a good gaming experience. In this context, network providers are encouraged to implement QoE-aware management systems to guarantee a desired Quality of Experience (QoE), in which Machine Learning (ML) models achieve the state-of-the-art on QoE estimation/monitoring. However, it is hard to create ML models that generalize to different contexts, especially since QoE perception is subjective and varies among games and players. This paper employs transfer learning and fine-tuning to adjust a source model to different target domains. First, we performed a subjective QoE assessment with real users playing on a realistic testbed. Based on this, we derived four datasets, one being the source dataset (to create the source model) and three distinct target datasets. Experiments show that transfer learning can decrease the average MSE error by at least 41.6% compared to the source model performance on the target datasets while decreasing the demand for labeled data by at least 81.1%. Furthermore, the improvement is greater when compared to models trained from scratch for each target dataset.
{"title":"Transfer Learning-Based QoE Estimation For Different Cloud Gaming Contexts","authors":"Marcos Carvalho, Daniel Soares, D. Macedo","doi":"10.1109/NetSoft57336.2023.10175441","DOIUrl":"https://doi.org/10.1109/NetSoft57336.2023.10175441","url":null,"abstract":"Cloud Gaming renders game data in the cloud and forwards it to players over the network. While this reduces hardware costs for players, it introduces challenges in network management and delivering a good gaming experience. In this context, network providers are encouraged to implement QoE-aware management systems to guarantee a desired Quality of Experience (QoE), in which Machine Learning (ML) models achieve the state-of-the-art on QoE estimation/monitoring. However, it is hard to create ML models that generalize to different contexts, especially since QoE perception is subjective and varies among games and players. This paper employs transfer learning and fine-tuning to adjust a source model to different target domains. First, we performed a subjective QoE assessment with real users playing on a realistic testbed. Based on this, we derived four datasets, one being the source dataset (to create the source model) and three distinct target datasets. Experiments show that transfer learning can decrease the average MSE error by at least 41.6% compared to the source model performance on the target datasets while decreasing the demand for labeled data by at least 81.1%. Furthermore, the improvement is greater when compared to models trained from scratch for each target dataset.","PeriodicalId":223208,"journal":{"name":"2023 IEEE 9th International Conference on Network Softwarization (NetSoft)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121609793","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：