Pub Date : 2021-11-01DOI: 10.1109/ICNP52444.2021.9651970
Hao Du, S. Leng, Jianhua He, Longyu Zhou
Vehicle platooning is one of the advanced driving applications expected to be supported by the 5G vehicle to everything (V2X) communications. It holds great potentials on improving road efficiency, driving safety and fuel efficiency. Apart from the organization and internal communication of the platoons, real-time prediction of surrounding road users (such as vehicles and cyclists) is another critical issue. While artificial intelligence (AI) is receiving increasing interests on its application to trajectory prediction, there is a potential problem that the pre-trained neural network models may not well fit the current driving environment and needs online fine-tuning to maintain an acceptable high prediction accuracy. In this paper, we propose a digital twin based real-time trajectory prediction scheme for platoons of connected intelligent vehicles. In this scheme the head vehicle of a platoon senses the surrounding vehicles. A LSTM neural network is applied for real-time trajectory prediction with the sensing outcomes. The head vehicle controls the offloading of the trajectory data and maintains a digital twin to optimize the update of LSTM model. In the digital twin a Deep-Q Learning (DQN) algorithm is utilized for adaptive fine tuning of the LSTM model, to ensure the prediction accuracy and minimize the consumption of communication and computing resources. A real-world dataset is developed from the KITTI datasets for simulations. The simulation results show that the proposed trajectory prediction scheme can maintain a prediction accuracy for safe platooning and reduce the delay of updating the neural networks by up to 40%.
{"title":"Digital Twin Based Trajectory Prediction for Platoons of Connected Intelligent Vehicles","authors":"Hao Du, S. Leng, Jianhua He, Longyu Zhou","doi":"10.1109/ICNP52444.2021.9651970","DOIUrl":"https://doi.org/10.1109/ICNP52444.2021.9651970","url":null,"abstract":"Vehicle platooning is one of the advanced driving applications expected to be supported by the 5G vehicle to everything (V2X) communications. It holds great potentials on improving road efficiency, driving safety and fuel efficiency. Apart from the organization and internal communication of the platoons, real-time prediction of surrounding road users (such as vehicles and cyclists) is another critical issue. While artificial intelligence (AI) is receiving increasing interests on its application to trajectory prediction, there is a potential problem that the pre-trained neural network models may not well fit the current driving environment and needs online fine-tuning to maintain an acceptable high prediction accuracy. In this paper, we propose a digital twin based real-time trajectory prediction scheme for platoons of connected intelligent vehicles. In this scheme the head vehicle of a platoon senses the surrounding vehicles. A LSTM neural network is applied for real-time trajectory prediction with the sensing outcomes. The head vehicle controls the offloading of the trajectory data and maintains a digital twin to optimize the update of LSTM model. In the digital twin a Deep-Q Learning (DQN) algorithm is utilized for adaptive fine tuning of the LSTM model, to ensure the prediction accuracy and minimize the consumption of communication and computing resources. A real-world dataset is developed from the KITTI datasets for simulations. The simulation results show that the proposed trajectory prediction scheme can maintain a prediction accuracy for safe platooning and reduce the delay of updating the neural networks by up to 40%.","PeriodicalId":343813,"journal":{"name":"2021 IEEE 29th International Conference on Network Protocols (ICNP)","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126759645","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 : 2021-11-01DOI: 10.1109/ICNP52444.2021.9651945
José Álamos, Peter Kietzmann, T. Schmidt, Matthias Wählisch
LoRa is a popular technology that enables long-range wireless communication (kilometers) at low energy consumption. The transmission exhibits low throughput and underlies duty cycle restrictions. Long on-air times (up to seconds) and range are susceptible to interference. In parallel, common LoRa-devices are battery driven and should mainly sleep. LoRaWAN is the system that defines the LoRa PHY, MAC, and a complete vertical stack. To deal with the above limitations, LoRaWAN imposes rigorous constraints, namely, a centralized network architecture that organizes media access, and heavily reduced downlink capacity. This makes it unusable for many deployments, control systems in particular. In this work, we combine IEEE802.15.4 DSME and LoRa to facilitate node-to-node communication. We present a DSME-LoRa mapping scheme and contribute a simulation model for validating new LoRa use-cases. Our results show 100% packet delivery and predictable latencies irrespective of network size.
{"title":"Poster: DSME-LoRa – A Flexible MAC for LoRa","authors":"José Álamos, Peter Kietzmann, T. Schmidt, Matthias Wählisch","doi":"10.1109/ICNP52444.2021.9651945","DOIUrl":"https://doi.org/10.1109/ICNP52444.2021.9651945","url":null,"abstract":"LoRa is a popular technology that enables long-range wireless communication (kilometers) at low energy consumption. The transmission exhibits low throughput and underlies duty cycle restrictions. Long on-air times (up to seconds) and range are susceptible to interference. In parallel, common LoRa-devices are battery driven and should mainly sleep. LoRaWAN is the system that defines the LoRa PHY, MAC, and a complete vertical stack. To deal with the above limitations, LoRaWAN imposes rigorous constraints, namely, a centralized network architecture that organizes media access, and heavily reduced downlink capacity. This makes it unusable for many deployments, control systems in particular. In this work, we combine IEEE802.15.4 DSME and LoRa to facilitate node-to-node communication. We present a DSME-LoRa mapping scheme and contribute a simulation model for validating new LoRa use-cases. Our results show 100% packet delivery and predictable latencies irrespective of network size.","PeriodicalId":343813,"journal":{"name":"2021 IEEE 29th International Conference on Network Protocols (ICNP)","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123241743","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 : 2021-11-01DOI: 10.1109/ICNP52444.2021.9651972
Natchanon Luangsomboon, J. Liebeherr
Hierarchical link sharing addresses the demand for fine-grain traffic control at multiple levels of aggregation. At present, packet schedulers that can support hierarchical link sharing are not suitable for an implementation at line rates, whereas deployed schedulers perform poorly at distributing excess capacity to classes that need additional bandwidth. We present HLS, a packet scheduler that ensures a hierarchical max-min fair allocation of the link bandwidth. HLS supports minimum rate guarantees and isolation between classes. Since it is realized as a non-hierarchical round-robin scheduler, it is suitable to operate at high rates. We implement HLS in the Linux kernel and evaluate it with respect to achieved rate allocations and overhead. We compare the results with those obtained for CBQ and HTB, the existing scheduling algorithms in Linux for hierarchical link sharing. We show that the overhead of HLS is comparable to that of other classful packet schedulers.
{"title":"HLS: A Packet Scheduler for Hierarchical Fairness","authors":"Natchanon Luangsomboon, J. Liebeherr","doi":"10.1109/ICNP52444.2021.9651972","DOIUrl":"https://doi.org/10.1109/ICNP52444.2021.9651972","url":null,"abstract":"Hierarchical link sharing addresses the demand for fine-grain traffic control at multiple levels of aggregation. At present, packet schedulers that can support hierarchical link sharing are not suitable for an implementation at line rates, whereas deployed schedulers perform poorly at distributing excess capacity to classes that need additional bandwidth. We present HLS, a packet scheduler that ensures a hierarchical max-min fair allocation of the link bandwidth. HLS supports minimum rate guarantees and isolation between classes. Since it is realized as a non-hierarchical round-robin scheduler, it is suitable to operate at high rates. We implement HLS in the Linux kernel and evaluate it with respect to achieved rate allocations and overhead. We compare the results with those obtained for CBQ and HTB, the existing scheduling algorithms in Linux for hierarchical link sharing. We show that the overhead of HLS is comparable to that of other classful packet schedulers.","PeriodicalId":343813,"journal":{"name":"2021 IEEE 29th International Conference on Network Protocols (ICNP)","volume":"155 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116034426","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 : 2021-11-01DOI: 10.1109/ICNP52444.2021.9651909
S. Bhatti, Gregor Haywood, Ryo Yanagida
We describe protocol features to provide both Identity Privacy and Location Privacy at the network layer that are truly end-to-end, strengthening the trust model by constraining the boundary of trust to only the communicating parties. We show that Identity Privacy and Location Privacy can be provided by changing only the addressing model, whilst still remaining compatible with IPv6. Using the Identifier-Locator Network Protocol (ILNP), it is possible to use ephemeral end-system ILNP Node Identity (NID) values to improve identity privacy. Using the ILNP Locator values with dynamic bindings, it is possible to use multiple IPv6 routing prefixes as network Locator (L64) values to provide (topological) location privacy. This is achieved: (a) whilst maintaining end-to-end state for transport protocols, without proxies, tunnels, or gateways at the transport layer or application layer; and (b) without the use of cryptographic techniques, so performance is not impacted.
{"title":"End-to-End Privacy for Identity & Location with IP","authors":"S. Bhatti, Gregor Haywood, Ryo Yanagida","doi":"10.1109/ICNP52444.2021.9651909","DOIUrl":"https://doi.org/10.1109/ICNP52444.2021.9651909","url":null,"abstract":"We describe protocol features to provide both Identity Privacy and Location Privacy at the network layer that are truly end-to-end, strengthening the trust model by constraining the boundary of trust to only the communicating parties. We show that Identity Privacy and Location Privacy can be provided by changing only the addressing model, whilst still remaining compatible with IPv6. Using the Identifier-Locator Network Protocol (ILNP), it is possible to use ephemeral end-system ILNP Node Identity (NID) values to improve identity privacy. Using the ILNP Locator values with dynamic bindings, it is possible to use multiple IPv6 routing prefixes as network Locator (L64) values to provide (topological) location privacy. This is achieved: (a) whilst maintaining end-to-end state for transport protocols, without proxies, tunnels, or gateways at the transport layer or application layer; and (b) without the use of cryptographic techniques, so performance is not impacted.","PeriodicalId":343813,"journal":{"name":"2021 IEEE 29th International Conference on Network Protocols (ICNP)","volume":"322 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116121794","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 : 2021-11-01DOI: 10.1109/ICNP52444.2021.9651910
Jie Zhang, Chuan-gui Ma, W. Wang, Kai Zheng, Yong Cui
The main iteration goal of transport protocols is to optimize the performance of specific modules. In this poster, we propose a framework named EasyTrans, that enables fast iteration of transport protocol modules. With EasyTrans, developers can focus on the modules they want to iterate and no longer need to deal with other unnecessary parts of the transport protocol. Through different module calling modes, EasyTrans enables high performance even if the modules use algorithms that require sophisticated computation such as machine learning. We implement EasyTrans based on QUIC. Evaluation results show that the overhead of EasyTrans is slight.
{"title":"Poster: EasyTrans: Enable Fast Iteration of Transport Protocol","authors":"Jie Zhang, Chuan-gui Ma, W. Wang, Kai Zheng, Yong Cui","doi":"10.1109/ICNP52444.2021.9651910","DOIUrl":"https://doi.org/10.1109/ICNP52444.2021.9651910","url":null,"abstract":"The main iteration goal of transport protocols is to optimize the performance of specific modules. In this poster, we propose a framework named EasyTrans, that enables fast iteration of transport protocol modules. With EasyTrans, developers can focus on the modules they want to iterate and no longer need to deal with other unnecessary parts of the transport protocol. Through different module calling modes, EasyTrans enables high performance even if the modules use algorithms that require sophisticated computation such as machine learning. We implement EasyTrans based on QUIC. Evaluation results show that the overhead of EasyTrans is slight.","PeriodicalId":343813,"journal":{"name":"2021 IEEE 29th International Conference on Network Protocols (ICNP)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125209451","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 : 2021-11-01DOI: 10.1109/ICNP52444.2021.9651960
Fan Yang, Zerui Tian
In-network caching is a basic feature of ICN architecture. Traditional ICN is distributed, which means the locations of content blocks cannot be adjusted precisely. Therefore, the cache allocation in traditional ICN is hard to approach optimization. With the aid of centralized controllers provided by SDN, ICN can manipulate the cache allocation with high flexibility. Heuristic algorithms have been applied to the cache allocation of ICN with centralized controllers but cannot guarantee the feasibility of solutions because of the feature of randomness. This paper proposes a caching strategy named MRPGA based on genetic algorithms. The mechanism of MRPGA guarantees the feasibility of solutions and accelerates convergence. Also, the simulations show that MRPGA figures out a better cache distribution in a shorter time than the genetic algorithm.
{"title":"MRPGA: A Genetic-Algorithm-based In-network Caching for Information-Centric Networking","authors":"Fan Yang, Zerui Tian","doi":"10.1109/ICNP52444.2021.9651960","DOIUrl":"https://doi.org/10.1109/ICNP52444.2021.9651960","url":null,"abstract":"In-network caching is a basic feature of ICN architecture. Traditional ICN is distributed, which means the locations of content blocks cannot be adjusted precisely. Therefore, the cache allocation in traditional ICN is hard to approach optimization. With the aid of centralized controllers provided by SDN, ICN can manipulate the cache allocation with high flexibility. Heuristic algorithms have been applied to the cache allocation of ICN with centralized controllers but cannot guarantee the feasibility of solutions because of the feature of randomness. This paper proposes a caching strategy named MRPGA based on genetic algorithms. The mechanism of MRPGA guarantees the feasibility of solutions and accelerates convergence. Also, the simulations show that MRPGA figures out a better cache distribution in a shorter time than the genetic algorithm.","PeriodicalId":343813,"journal":{"name":"2021 IEEE 29th International Conference on Network Protocols (ICNP)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125627146","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 : 2021-11-01DOI: 10.1109/ICNP52444.2021.9651927
Lei Du, R. Huo
In the Industrial Internet of Things (IIoT) scenario, the increase of surveillance equipment brings challenges to the transmission of real-time video. It needs more efficient approaches to finish video transmission with more stability and accuracy. Therefore, we propose a self-adaptive transmission scheme of videos for multi-capture terminals under IIoT in this paper. To fit for the constant variation of network environment, we compress the videos that wait for transmitting from multi-capture terminals by reducing the non-key frames with Graph Convolutional Network (GCN). Moreover, a self-adaptive strategy of transmission is implemented on the Mobile Edge Computing (MEC) server to adjust the transmission volume of processed videos, and a multi-objective optimization algorithm is utilized to optimize the strategy of transmission during the video transmission. The relative experiments are conducted to validate the performance of the proposed scheme.
{"title":"Real-time Video Transmission Optimization Based on Edge Computing in IIoT","authors":"Lei Du, R. Huo","doi":"10.1109/ICNP52444.2021.9651927","DOIUrl":"https://doi.org/10.1109/ICNP52444.2021.9651927","url":null,"abstract":"In the Industrial Internet of Things (IIoT) scenario, the increase of surveillance equipment brings challenges to the transmission of real-time video. It needs more efficient approaches to finish video transmission with more stability and accuracy. Therefore, we propose a self-adaptive transmission scheme of videos for multi-capture terminals under IIoT in this paper. To fit for the constant variation of network environment, we compress the videos that wait for transmitting from multi-capture terminals by reducing the non-key frames with Graph Convolutional Network (GCN). Moreover, a self-adaptive strategy of transmission is implemented on the Mobile Edge Computing (MEC) server to adjust the transmission volume of processed videos, and a multi-objective optimization algorithm is utilized to optimize the strategy of transmission during the video transmission. The relative experiments are conducted to validate the performance of the proposed scheme.","PeriodicalId":343813,"journal":{"name":"2021 IEEE 29th International Conference on Network Protocols (ICNP)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131377491","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 : 2021-11-01DOI: 10.1109/ICNP52444.2021.9651985
Qian Chen, Jiliang Wang
LoRa has been shown as a promising Low-Power Wide Area Network (LPWAN) technology to connect millions of devices for the Internet of Things by providing long-distance low-power communication in a very low SNR. Real LoRa networks, however, suffer from severe packet collisions. Existing collision resolution approaches introduce a high SNR loss, i.e., require a much higher SNR than LoRa. To push the limit of LoRa collision decoding, we present AlignTrack, the first LoRa collision decoding approach that can work in the SNR limit of the original LoRa. Our key finding is that a LoRa chirp aligned with a decoding window should lead to the highest peak in the frequency domain and thus has the least SNR loss. By aligning a moving window with different packets, we separate packets by identifying the aligned chirp in each window. We theoretically prove this leads to the minimal SNR loss. In practical implementation, we address two key challenges: (1) accurately detecting the start of each packet, and (2) separating collided packets in each window in the presence of CFO and inter-packet interference. We implement AlignTrack on HackRF One and compare its performance with the state-of-the-arts. The evaluation results show that AlignTrack improves network throughput by 1.68× compared with NScale and 3× compared with CoLoRa.
{"title":"AlignTrack: Push the Limit of LoRa Collision Decoding","authors":"Qian Chen, Jiliang Wang","doi":"10.1109/ICNP52444.2021.9651985","DOIUrl":"https://doi.org/10.1109/ICNP52444.2021.9651985","url":null,"abstract":"LoRa has been shown as a promising Low-Power Wide Area Network (LPWAN) technology to connect millions of devices for the Internet of Things by providing long-distance low-power communication in a very low SNR. Real LoRa networks, however, suffer from severe packet collisions. Existing collision resolution approaches introduce a high SNR loss, i.e., require a much higher SNR than LoRa. To push the limit of LoRa collision decoding, we present AlignTrack, the first LoRa collision decoding approach that can work in the SNR limit of the original LoRa. Our key finding is that a LoRa chirp aligned with a decoding window should lead to the highest peak in the frequency domain and thus has the least SNR loss. By aligning a moving window with different packets, we separate packets by identifying the aligned chirp in each window. We theoretically prove this leads to the minimal SNR loss. In practical implementation, we address two key challenges: (1) accurately detecting the start of each packet, and (2) separating collided packets in each window in the presence of CFO and inter-packet interference. We implement AlignTrack on HackRF One and compare its performance with the state-of-the-arts. The evaluation results show that AlignTrack improves network throughput by 1.68× compared with NScale and 3× compared with CoLoRa.","PeriodicalId":343813,"journal":{"name":"2021 IEEE 29th International Conference on Network Protocols (ICNP)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124665613","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 : 2021-11-01DOI: 10.1109/ICNP52444.2021.9651971
Lei Zhang, Yongchang Cui, Junchen Pan, Yong Jiang
The deadline requirements of real-time applications rapidly increase in recent years (e.g., cloud gaming, cloud VR, online conferencing). Due to diverse network conditions, meeting deadline requirements for these applications has become one of the research hotspots. However, the current schemes focus on providing high bitrate instead of meeting deadline requirements. In this paper, we propose D3T, a flexible deadline-aware transmission mechanism that aims to improve user quality of experience (QoE) for real-time video streaming. To fulfill the diverse deadline requirements over fluctuating network conditions, D3T uses a deadline-aware scheduler to select the high priority frame before the deadline. To reduce congestion and retransmission delay, we leverage a deep reinforcement learning algorithm to make decisions of sending rate and FEC (forward error correction) redundancy ratio based on observed network status and frame information. We evaluate D3T via trace-driven simulator spanning diverse network environments, video contents and QoE metrics. D3T significantly improves the frame completion rate by reducing the bandwidth waste before the deadline. In the considered scenarios, D3T outperforms previously approaches with the improvements in average QoE of 57%.
{"title":"Deadline-Aware Transmission Control for Real-Time Video Streaming","authors":"Lei Zhang, Yongchang Cui, Junchen Pan, Yong Jiang","doi":"10.1109/ICNP52444.2021.9651971","DOIUrl":"https://doi.org/10.1109/ICNP52444.2021.9651971","url":null,"abstract":"The deadline requirements of real-time applications rapidly increase in recent years (e.g., cloud gaming, cloud VR, online conferencing). Due to diverse network conditions, meeting deadline requirements for these applications has become one of the research hotspots. However, the current schemes focus on providing high bitrate instead of meeting deadline requirements. In this paper, we propose D3T, a flexible deadline-aware transmission mechanism that aims to improve user quality of experience (QoE) for real-time video streaming. To fulfill the diverse deadline requirements over fluctuating network conditions, D3T uses a deadline-aware scheduler to select the high priority frame before the deadline. To reduce congestion and retransmission delay, we leverage a deep reinforcement learning algorithm to make decisions of sending rate and FEC (forward error correction) redundancy ratio based on observed network status and frame information. We evaluate D3T via trace-driven simulator spanning diverse network environments, video contents and QoE metrics. D3T significantly improves the frame completion rate by reducing the bandwidth waste before the deadline. In the considered scenarios, D3T outperforms previously approaches with the improvements in average QoE of 57%.","PeriodicalId":343813,"journal":{"name":"2021 IEEE 29th International Conference on Network Protocols (ICNP)","volume":"51 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125172627","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：