{"title":"Attention-based LSTM for Controller Load Prediction in Software-Defined Networks✱","authors":"Yong Liu, Quanze Liu, Qian Meng","doi":"10.1145/3600061.3603124","DOIUrl":"https://doi.org/10.1145/3600061.3603124","url":null,"abstract":"","PeriodicalId":228934,"journal":{"name":"Proceedings of the 7th Asia-Pacific Workshop on Networking","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117289099","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}
Reactive congestion control (RCC) protocols have undergone decades of evolution, where senders first send data packets and then back off when congestion occurs. Recently, there has been a surge of interest in proactive congestion control (PCC) that allocates bandwidth before transmission. Despite its potential, we found that there are certain scenarios where PCC may fall short. In this paper, we aim to provide a comprehensive understanding of PCC and motivate further exploration of this area. We conduct case studies and leverage NS3 simulations to compare state-of-the-art PCC with RCCs, delving into the real dilemma of PCC.
{"title":"Dilemma of Proactive Congestion Control Protocols","authors":"Kexin Liu, Chen Tian, Xiaoliang Wang, Wanchun Dou, Guihai Chen","doi":"10.1145/3600061.3603123","DOIUrl":"https://doi.org/10.1145/3600061.3603123","url":null,"abstract":"Reactive congestion control (RCC) protocols have undergone decades of evolution, where senders first send data packets and then back off when congestion occurs. Recently, there has been a surge of interest in proactive congestion control (PCC) that allocates bandwidth before transmission. Despite its potential, we found that there are certain scenarios where PCC may fall short. In this paper, we aim to provide a comprehensive understanding of PCC and motivate further exploration of this area. We conduct case studies and leverage NS3 simulations to compare state-of-the-art PCC with RCCs, delving into the real dilemma of PCC.","PeriodicalId":228934,"journal":{"name":"Proceedings of the 7th Asia-Pacific Workshop on Networking","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123908712","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}
Online recommendation systems play critical roles in enhancing user experience by helping them find the most interesting videos from a vast amount of content. However, the existing recommendation modules and video transmission modules in the industry often operate independently, resulting in the recommendation model providing some videos that cannot be transmitted within the specified deadlines successfully. This can lead to an inferior watching experience for users and resource waste for video providers. To address this, we propose a novel framework called NetRec, which for the first time optimizes the recommendation quality by jointly considering the network transmission. We accomplish this by re-ranking the top-N videos obtained from the recommendation system and selecting the top-M (M is approximately half of N) videos that provide the maximum overall revenue, e.g., video playing time while considering the network status. The entire system comprises network measurement, video quality estimation, and multi-objective optimization modules. Real-world Internet results show that our framework can increase users’ video playing time by 20% to 160%. Furthermore, we provide several promising directions for further improving the video recommendation quality under our NetRec framework, which jointly considers the network for the recommendation.
{"title":"Beyond the Content: Considering the Network for Online Video Recommendation","authors":"Lihui Lang, Meiqi Hu, Changhua Pei, Guo Chen","doi":"10.1145/3600061.3600075","DOIUrl":"https://doi.org/10.1145/3600061.3600075","url":null,"abstract":"Online recommendation systems play critical roles in enhancing user experience by helping them find the most interesting videos from a vast amount of content. However, the existing recommendation modules and video transmission modules in the industry often operate independently, resulting in the recommendation model providing some videos that cannot be transmitted within the specified deadlines successfully. This can lead to an inferior watching experience for users and resource waste for video providers. To address this, we propose a novel framework called NetRec, which for the first time optimizes the recommendation quality by jointly considering the network transmission. We accomplish this by re-ranking the top-N videos obtained from the recommendation system and selecting the top-M (M is approximately half of N) videos that provide the maximum overall revenue, e.g., video playing time while considering the network status. The entire system comprises network measurement, video quality estimation, and multi-objective optimization modules. Real-world Internet results show that our framework can increase users’ video playing time by 20% to 160%. Furthermore, we provide several promising directions for further improving the video recommendation quality under our NetRec framework, which jointly considers the network for the recommendation.","PeriodicalId":228934,"journal":{"name":"Proceedings of the 7th Asia-Pacific Workshop on Networking","volume":"150 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131562332","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}
With the continuously increasing scale of deep neural network models, there is a clear trend towards distributed DNN model training. State-of-the-art training frameworks support this approach using collective communication libraries such as NCCL, MPI, Gloo, and Horovod. These libraries have many parameters that can be adjusted to fit different hardware environments, and these parameters can greatly impact training performance. Therefore, careful tuning of parameters for each training environment is required. However, given the large parameter space, manual exploration can be time-consuming and laborious. In this poster, we introduce AFNFA, which stands for AI For Network For AI. It is an automated program that utilizes machine learning and simulated annealing to explore NCCL parameters. Preliminary evaluation results demonstrate that compared to the default configuration, the configuration explored by AFNFA improves NCCL communication performance by 22.90%.
随着深度神经网络模型规模的不断扩大，分布式DNN模型训练有明显的趋势。最先进的培训框架使用集体通信库(如NCCL、MPI、Gloo和Horovod)支持这种方法。这些库有许多参数，可以调整以适应不同的硬件环境，这些参数可以极大地影响训练性能。因此，需要仔细调整每个训练环境的参数。然而，考虑到大的参数空间，人工探索可能是费时费力的。在这张海报中，我们介绍了AFNFA，即AI for Network for AI。它是一个自动化程序，利用机器学习和模拟退火来探索NCCL参数。初步评估结果表明，与默认配置相比，AFNFA探索的配置使NCCL通信性能提高了22.90%。
{"title":"AFNFA: An Approach to Automate NCCL Configuration Exploration","authors":"Zibo Wang, Yuhang Zhou, Chen Tian, Xiaoliang Wang, Xianping Chen","doi":"10.1145/3600061.3600068","DOIUrl":"https://doi.org/10.1145/3600061.3600068","url":null,"abstract":"With the continuously increasing scale of deep neural network models, there is a clear trend towards distributed DNN model training. State-of-the-art training frameworks support this approach using collective communication libraries such as NCCL, MPI, Gloo, and Horovod. These libraries have many parameters that can be adjusted to fit different hardware environments, and these parameters can greatly impact training performance. Therefore, careful tuning of parameters for each training environment is required. However, given the large parameter space, manual exploration can be time-consuming and laborious. In this poster, we introduce AFNFA, which stands for AI For Network For AI. It is an automated program that utilizes machine learning and simulated annealing to explore NCCL parameters. Preliminary evaluation results demonstrate that compared to the default configuration, the configuration explored by AFNFA improves NCCL communication performance by 22.90%.","PeriodicalId":228934,"journal":{"name":"Proceedings of the 7th Asia-Pacific Workshop on Networking","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128424988","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}
Young Choi, Jun-Sup Yoon, YoungGyoun Moon, KyoungSoo Park
The Maximum Transmission Unit (MTU) refers to the largest packet size that can be transferred on a particular layer-3 network. As the dominance of Ethernet prevails, the "de-facto" standard MTU of 1500B has become universal in the wide-area networks. Unfortunately, the current MTU size overly limits the transmission performance especially when the underlying link speed rapidly increases while the CPU advancement stagnates. In this work, we investigate the potential impact of large MTU on fast-growing cellular core networks. First, we analyze the performance trend over the different MTU sizes on endpoint receivers as well as on User Plane Function (UPF) in a cellular core network that handles all data packets. Second, we present our dynamic MTU translation technique to transparently apply a large MTU inside a cellular core network without requiring update on other networks in the Internet. We observe that that large MTU is beneficial to both traffic endpoints and UPF, and our evaluation shows that dynamic packet merging scales the UPF performance by up to 4.9x, reaching 628 Gbps with only eight CPU cores.
最大传输单元(Maximum Transmission Unit, MTU)是指在特定的三层网络中可以传输的最大数据包大小。随着以太网的盛行，1500B的“事实”标准MTU在广域网中已成为通用标准。不幸的是，当前MTU的大小过度限制了传输性能，特别是当底层链路速度快速增加而CPU进度停滞时。在这项工作中，我们研究了大MTU对快速增长的蜂窝核心网的潜在影响。首先，我们分析了端点接收器上不同MTU大小的性能趋势，以及处理所有数据包的蜂窝核心网络中的用户平面功能(UPF)。其次，我们提出了动态MTU转换技术，以透明地在蜂窝核心网中应用大型MTU，而无需在互联网上的其他网络上进行更新。我们观察到，大的MTU对流量端点和UPF都是有益的，我们的评估表明，动态分组合并将UPF性能提高了4.9倍，仅用8个CPU内核就达到628 Gbps。
{"title":"Is Large MTU Beneficial to Cellular Core Networks?","authors":"Young Choi, Jun-Sup Yoon, YoungGyoun Moon, KyoungSoo Park","doi":"10.1145/3600061.3600081","DOIUrl":"https://doi.org/10.1145/3600061.3600081","url":null,"abstract":"The Maximum Transmission Unit (MTU) refers to the largest packet size that can be transferred on a particular layer-3 network. As the dominance of Ethernet prevails, the \"de-facto\" standard MTU of 1500B has become universal in the wide-area networks. Unfortunately, the current MTU size overly limits the transmission performance especially when the underlying link speed rapidly increases while the CPU advancement stagnates. In this work, we investigate the potential impact of large MTU on fast-growing cellular core networks. First, we analyze the performance trend over the different MTU sizes on endpoint receivers as well as on User Plane Function (UPF) in a cellular core network that handles all data packets. Second, we present our dynamic MTU translation technique to transparently apply a large MTU inside a cellular core network without requiring update on other networks in the Internet. We observe that that large MTU is beneficial to both traffic endpoints and UPF, and our evaluation shows that dynamic packet merging scales the UPF performance by up to 4.9x, reaching 628 Gbps with only eight CPU cores.","PeriodicalId":228934,"journal":{"name":"Proceedings of the 7th Asia-Pacific Workshop on Networking","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130460615","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}
Computing power network (CPN) has been proposed to allocate and schedule computing power resources among cloud, network, and edge according to the needs of computing services. CPN can improve the utilization rate of various computing resource pools. However, it brings another challenge that how to ensure the security of multi-tenant information and the resource information which they rent. To solve this problem, we propose an isolation architecture in CPN, add a tenant mapping management module in network control layer firstly. Then we design the security mapping process between the tenant and the computing routing node based on this architecture. At last, we propose a mapping method between tenants and computing routing nodes based on hash ring which can avoid the problem of data migration caused by increasing the number of computing routing nodes. In the future, we will study the mapping algorithm to improve the efficiency of CPN.
{"title":"A Security Mapping Approach between Multi Tenant and Computing Routing Nodes in CPN","authors":"Jiacong Li, Hang Lv, Bo Lei, Yunpeng Xie","doi":"10.1145/3600061.3603121","DOIUrl":"https://doi.org/10.1145/3600061.3603121","url":null,"abstract":"Computing power network (CPN) has been proposed to allocate and schedule computing power resources among cloud, network, and edge according to the needs of computing services. CPN can improve the utilization rate of various computing resource pools. However, it brings another challenge that how to ensure the security of multi-tenant information and the resource information which they rent. To solve this problem, we propose an isolation architecture in CPN, add a tenant mapping management module in network control layer firstly. Then we design the security mapping process between the tenant and the computing routing node based on this architecture. At last, we propose a mapping method between tenants and computing routing nodes based on hash ring which can avoid the problem of data migration caused by increasing the number of computing routing nodes. In the future, we will study the mapping algorithm to improve the efficiency of CPN.","PeriodicalId":228934,"journal":{"name":"Proceedings of the 7th Asia-Pacific Workshop on Networking","volume":"81 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126413689","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}
The essence of the User Plane Function (UPF) is strong forwarding, and virtualisation architecture by software implemented can not meet high-performance requirements. Thus, hardware acceleration becomes an option. However, existing offloading schemes as an accelerator are not prominent in the latency. In this paper, we implement a demo called cUPFCard to offload UPF into a smart NIC based on FPGA platform, which can provide higher throughput with lower latency. Experiments show that cUPFCard is feasible in the real network. Moreover, the throughput is improved 24 times, and the latency is decreased 41 times.
{"title":"cUPFCard: High-Performance User Plane Function based on FPGA","authors":"Cong Zhou, Baokang Zhao, Baosheng Wang","doi":"10.1145/3600061.3603119","DOIUrl":"https://doi.org/10.1145/3600061.3603119","url":null,"abstract":"The essence of the User Plane Function (UPF) is strong forwarding, and virtualisation architecture by software implemented can not meet high-performance requirements. Thus, hardware acceleration becomes an option. However, existing offloading schemes as an accelerator are not prominent in the latency. In this paper, we implement a demo called cUPFCard to offload UPF into a smart NIC based on FPGA platform, which can provide higher throughput with lower latency. Experiments show that cUPFCard is feasible in the real network. Moreover, the throughput is improved 24 times, and the latency is decreased 41 times.","PeriodicalId":228934,"journal":{"name":"Proceedings of the 7th Asia-Pacific Workshop on Networking","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134637911","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}
Mubashir Anwar, Fangping Lan, Anduo Wang, Matthew Caesar
The future of static verification in networking may be obscured by two clouds: the complexity of distributed systems with highly concurrent events, and the decision-making on infrastructures growing without a premeditated plan. This poster discusses a possible solution to these issues, in which the huge space of analyzing distributed systems and the macro-questions of system evolution are addressed by a common structure, a logical implication problem which we call indirect troubleshooting. The usefulness and feasibility of indirect troubleshooting is illustrated by a preliminary realization with the chase, a remarkable process for mechanically deciding implications.
{"title":"Indirect Network Troubleshooting with The Chase","authors":"Mubashir Anwar, Fangping Lan, Anduo Wang, Matthew Caesar","doi":"10.1145/3600061.3603137","DOIUrl":"https://doi.org/10.1145/3600061.3603137","url":null,"abstract":"The future of static verification in networking may be obscured by two clouds: the complexity of distributed systems with highly concurrent events, and the decision-making on infrastructures growing without a premeditated plan. This poster discusses a possible solution to these issues, in which the huge space of analyzing distributed systems and the macro-questions of system evolution are addressed by a common structure, a logical implication problem which we call indirect troubleshooting. The usefulness and feasibility of indirect troubleshooting is illustrated by a preliminary realization with the chase, a remarkable process for mechanically deciding implications.","PeriodicalId":228934,"journal":{"name":"Proceedings of the 7th Asia-Pacific Workshop on Networking","volume":"57 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133066262","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：