Pub Date : 2022-06-30DOI: 10.1109/ISCC55528.2022.9912988
Lucas N. Silva, Paulo H. L. Rettore, Vinícius F. S. Mota, B. P. Santos
Due to the increasing location-aware devices, mobility traces datasets have become an essential source for smart cities planning. Given this scenario, we propose MobVis, a framework to characterize mobility traces through different metrics, allowing comparisons between different mobility traces in a simplified way. Furthermore, MobVis can extract and visualize spatial, temporal, and social aspects of mobility data through a Web interface. MobVis architecture has five main components: input data; data preparation; data processing and analysis to extract mobility metrics; visualization; and a web interface. To demonstrate the framework's process, we created a use case analyzing the characteristics of two distinct traces (Taxi and IoT-Objects). Then, through different metrics, we evaluated the data in two aspects: i) descriptive, through a set of graphics and quantitative data that enables characterizing each trace; and ii) comparative, presenting the main differences between the traces.
{"title":"MobVis: A Framework for Analysis and Visualization of Mobility Traces","authors":"Lucas N. Silva, Paulo H. L. Rettore, Vinícius F. S. Mota, B. P. Santos","doi":"10.1109/ISCC55528.2022.9912988","DOIUrl":"https://doi.org/10.1109/ISCC55528.2022.9912988","url":null,"abstract":"Due to the increasing location-aware devices, mobility traces datasets have become an essential source for smart cities planning. Given this scenario, we propose MobVis, a framework to characterize mobility traces through different metrics, allowing comparisons between different mobility traces in a simplified way. Furthermore, MobVis can extract and visualize spatial, temporal, and social aspects of mobility data through a Web interface. MobVis architecture has five main components: input data; data preparation; data processing and analysis to extract mobility metrics; visualization; and a web interface. To demonstrate the framework's process, we created a use case analyzing the characteristics of two distinct traces (Taxi and IoT-Objects). Then, through different metrics, we evaluated the data in two aspects: i) descriptive, through a set of graphics and quantitative data that enables characterizing each trace; and ii) comparative, presenting the main differences between the traces.","PeriodicalId":309606,"journal":{"name":"2022 IEEE Symposium on Computers and Communications (ISCC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127325576","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 : 2022-06-30DOI: 10.1109/ISCC55528.2022.9912965
Eftychia Lakka, George Hatzivasilis, Stylianos Karagiannis, Andreas D. Alexopoulos, M. Athanatos, S. Ioannidis, Manolis Chatzimpyrros, Grigoris Kalogiannis, G. Spanoudakis
Healthcare ecosystems form a critical type of infrastructures that provide valuable services in today societies. However, the underlying sensitive information is also of interest of malicious entities around the globe, with the attack volume being continuously increasing. Safeguarding this complex computerized setting constitutes a major challenge for the involved organizations. This paper presents an incident handling system for healthcare organizations and their supply-chain. The proposed approach utilizes swarm intelligence in order to assess the current security posture in a continuous basis and respond to attacks in real-time. The overall solution is based on the related NIST 800.61 standard and implements the operations of i) preparation, ii) detection and analysis, iii) containment, eradication, and recovery, and iv) post-incident activity. The system is developed under the EU funded project AI4HEALTHSEC and is applied in the relevant healthcare pilots.
{"title":"Incident Handling for Healthcare Organizations and Supply-Chains","authors":"Eftychia Lakka, George Hatzivasilis, Stylianos Karagiannis, Andreas D. Alexopoulos, M. Athanatos, S. Ioannidis, Manolis Chatzimpyrros, Grigoris Kalogiannis, G. Spanoudakis","doi":"10.1109/ISCC55528.2022.9912965","DOIUrl":"https://doi.org/10.1109/ISCC55528.2022.9912965","url":null,"abstract":"Healthcare ecosystems form a critical type of infrastructures that provide valuable services in today societies. However, the underlying sensitive information is also of interest of malicious entities around the globe, with the attack volume being continuously increasing. Safeguarding this complex computerized setting constitutes a major challenge for the involved organizations. This paper presents an incident handling system for healthcare organizations and their supply-chain. The proposed approach utilizes swarm intelligence in order to assess the current security posture in a continuous basis and respond to attacks in real-time. The overall solution is based on the related NIST 800.61 standard and implements the operations of i) preparation, ii) detection and analysis, iii) containment, eradication, and recovery, and iv) post-incident activity. The system is developed under the EU funded project AI4HEALTHSEC and is applied in the relevant healthcare pilots.","PeriodicalId":309606,"journal":{"name":"2022 IEEE Symposium on Computers and Communications (ISCC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125261489","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 : 2022-06-30DOI: 10.1109/ISCC55528.2022.9912781
I. Moursy, S. Ghanem, Mohamed Nazih ElDerini
Internet of Things (IoT) integrates the physical world with the Internet to facilitate sharing data among entities. These IoT applications bring security and management challenges. In this paper, using blockchain technology, an end-to-end secure architecture is proposed for an auditable tamper-proof log of sensors' data and events. A Role Based Access Control policy is enforced using smart contracts that controls access to both sensors' data and executing commands on actuators. For confidentiality, the data is kept encrypted in transit and at rest. In addition, the data is stored off-the-chain in a distributed content-based addressable network, while its address and access records are stored at the blockchain. The proposed design is suitable for real-time and mission-critical IoT applications. Finally, a proof-of-concept implementation shows the efficiency and scalability of the proposed architecture.
物联网(Internet of Things, IoT)将物理世界与互联网相结合，方便实体之间的数据共享。这些物联网应用带来了安全和管理方面的挑战。本文利用区块链技术，提出了一种端到端安全架构，用于传感器数据和事件的可审计防篡改日志。基于角色的访问控制策略使用智能合约来实施，智能合约控制对传感器数据的访问和对执行器执行命令。为了保密，数据在传输和静止时都是加密的。此外，数据存储在基于内容的分布式可寻址网络中，而其地址和访问记录存储在区块链中。提出的设计适用于实时和关键任务的物联网应用。最后，概念验证实现展示了所提出体系结构的效率和可扩展性。
{"title":"A Blockchain-Based Architecture for Access Control Management of IoT Applications","authors":"I. Moursy, S. Ghanem, Mohamed Nazih ElDerini","doi":"10.1109/ISCC55528.2022.9912781","DOIUrl":"https://doi.org/10.1109/ISCC55528.2022.9912781","url":null,"abstract":"Internet of Things (IoT) integrates the physical world with the Internet to facilitate sharing data among entities. These IoT applications bring security and management challenges. In this paper, using blockchain technology, an end-to-end secure architecture is proposed for an auditable tamper-proof log of sensors' data and events. A Role Based Access Control policy is enforced using smart contracts that controls access to both sensors' data and executing commands on actuators. For confidentiality, the data is kept encrypted in transit and at rest. In addition, the data is stored off-the-chain in a distributed content-based addressable network, while its address and access records are stored at the blockchain. The proposed design is suitable for real-time and mission-critical IoT applications. Finally, a proof-of-concept implementation shows the efficiency and scalability of the proposed architecture.","PeriodicalId":309606,"journal":{"name":"2022 IEEE Symposium on Computers and Communications (ISCC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125375403","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 : 2022-06-30DOI: 10.1109/ISCC55528.2022.9912881
Leandro M. Dallanora, A. G. Castro, R. I. T. D. C. Filho, F. Rossi, A. Lorenzon, M. C. Luizelli
In-band Network Telemetry (INT) is a novel net-work monitoring mechanism that improves fine-grained net-work visibility. Despite the increasing research efforts towards the orchestration of INT data acquisition, little has yet been done to efficiently collect telemetry data from the network considering monitoring applications requirements. In this paper, we introduce DyPro - a dynamic probing planning for INT. In particular, DyP ro ensures that telemetry dependencies are always satisfied by monitoring application requirements. We theoretically formalize it as a Mixed-Integer Linear Programming (MILP) optimization model and propose a heuristic procedure to efficiently solve it. Results show that DyP ro can outperform state-of-the-art solutions by up to 5x regarding the percentage of monitoring applications satisfied.
{"title":"DyPro: Dynamic Probing Planning for In-Band Network Telemetry","authors":"Leandro M. Dallanora, A. G. Castro, R. I. T. D. C. Filho, F. Rossi, A. Lorenzon, M. C. Luizelli","doi":"10.1109/ISCC55528.2022.9912881","DOIUrl":"https://doi.org/10.1109/ISCC55528.2022.9912881","url":null,"abstract":"In-band Network Telemetry (INT) is a novel net-work monitoring mechanism that improves fine-grained net-work visibility. Despite the increasing research efforts towards the orchestration of INT data acquisition, little has yet been done to efficiently collect telemetry data from the network considering monitoring applications requirements. In this paper, we introduce DyPro - a dynamic probing planning for INT. In particular, DyP ro ensures that telemetry dependencies are always satisfied by monitoring application requirements. We theoretically formalize it as a Mixed-Integer Linear Programming (MILP) optimization model and propose a heuristic procedure to efficiently solve it. Results show that DyP ro can outperform state-of-the-art solutions by up to 5x regarding the percentage of monitoring applications satisfied.","PeriodicalId":309606,"journal":{"name":"2022 IEEE Symposium on Computers and Communications (ISCC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126877327","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 : 2022-06-30DOI: 10.1109/ISCC55528.2022.9912909
M. Furini, E. Flisi
Social media and personal health might be a dan-gerous combination: people are influenced by what they read online and don't pay attention to who wrote what they read. What happened during the COVID-19 pandemic? Who were the opinion leaders on social media? What were the conversations about? How did the health institutions communicate? To under-stand this, we focus on Twitter, and we analyze more than three million of Italian-written tweets posted from January 2020 to December 2021. We propose a method to identify opinion leaders and to analyze the content of the conversations. Results show that: (i) opinion leaders are linked to what they say and when they say it; (ii) politicians, newscast, and ordinary people accounts were able to become opinion leaders during the pandemic; (iii) conversations moved from a medical focus (at the beginning of the pandemic) to a social focus (in the last months of 2021); (iv) absence of health care institutions among opinion leaders. These results show that our approach might be useful for those who want to monitor the social scenario in terms of health (e.g., to identify as soon as possible accounts against or critical to medicine or to health authorities).
{"title":"Opinion Leaders and Twitter: Metric Proposal and Psycholinguistic Analysis","authors":"M. Furini, E. Flisi","doi":"10.1109/ISCC55528.2022.9912909","DOIUrl":"https://doi.org/10.1109/ISCC55528.2022.9912909","url":null,"abstract":"Social media and personal health might be a dan-gerous combination: people are influenced by what they read online and don't pay attention to who wrote what they read. What happened during the COVID-19 pandemic? Who were the opinion leaders on social media? What were the conversations about? How did the health institutions communicate? To under-stand this, we focus on Twitter, and we analyze more than three million of Italian-written tweets posted from January 2020 to December 2021. We propose a method to identify opinion leaders and to analyze the content of the conversations. Results show that: (i) opinion leaders are linked to what they say and when they say it; (ii) politicians, newscast, and ordinary people accounts were able to become opinion leaders during the pandemic; (iii) conversations moved from a medical focus (at the beginning of the pandemic) to a social focus (in the last months of 2021); (iv) absence of health care institutions among opinion leaders. These results show that our approach might be useful for those who want to monitor the social scenario in terms of health (e.g., to identify as soon as possible accounts against or critical to medicine or to health authorities).","PeriodicalId":309606,"journal":{"name":"2022 IEEE Symposium on Computers and Communications (ISCC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127939413","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 : 2022-06-30DOI: 10.1109/ISCC55528.2022.9912866
Yan Wang, Q. Ren
Traffic prediction is a critical component of intel-ligent transportation systems. However, highly non-linear and dynamical spatial-temporal correlations propose challenges for traffic prediction, especially long-term prediction. We propose a spatial-temporal channel-attention based graph convolutional network (STCAGCN) to improve the accuracy of both long-term and short-term traffic flow prediction. Firstly we design an attention mechanism to learn complex temporal and spatial correlations. Then we develop the stacked spatial-temporal convo-lution layer to model complex temporal and spatial correlations. Each spatial-temporal convolution layer is composed of a gated time convolution network and a graph convolution network. We develop a gated time convolution network to model non-linear temporal correlations, which process long sequences through stacked dilated convolution. Moreover, the graph convolution network exploits the hidden spatial correlations via learning self-adaptive adjacency matrix. Experiment results on real-world datasets demonstrate that the proposed STCAGCN model obtains improvements over the state-of-the-art, especially for long-term traffic flow prediction.
{"title":"Dynamic Graph Convolutional Network for Long Short-term Traffic Flow Prediction","authors":"Yan Wang, Q. Ren","doi":"10.1109/ISCC55528.2022.9912866","DOIUrl":"https://doi.org/10.1109/ISCC55528.2022.9912866","url":null,"abstract":"Traffic prediction is a critical component of intel-ligent transportation systems. However, highly non-linear and dynamical spatial-temporal correlations propose challenges for traffic prediction, especially long-term prediction. We propose a spatial-temporal channel-attention based graph convolutional network (STCAGCN) to improve the accuracy of both long-term and short-term traffic flow prediction. Firstly we design an attention mechanism to learn complex temporal and spatial correlations. Then we develop the stacked spatial-temporal convo-lution layer to model complex temporal and spatial correlations. Each spatial-temporal convolution layer is composed of a gated time convolution network and a graph convolution network. We develop a gated time convolution network to model non-linear temporal correlations, which process long sequences through stacked dilated convolution. Moreover, the graph convolution network exploits the hidden spatial correlations via learning self-adaptive adjacency matrix. Experiment results on real-world datasets demonstrate that the proposed STCAGCN model obtains improvements over the state-of-the-art, especially for long-term traffic flow prediction.","PeriodicalId":309606,"journal":{"name":"2022 IEEE Symposium on Computers and Communications (ISCC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128036145","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 : 2022-06-30DOI: 10.1109/ISCC55528.2022.9913017
Rui Mao, Heming Ji, Di Cheng, Xiaoyu Wang, Yan Wang, Degang Sun
Most existing identity authentication technologies rely on some ways for the first login authentication, such as personal identification number (PIN), track, or biological characteristics. However, these ways exist plenty of security risks, which make people face password guessing attacks, trace attacks, and shoulder surfing attacks for a long time. Once the illegal users forge identity to complete authentication or bypass first login authentication, their subsequent behavior will become out of control. To solve the above problems, we propose an implicit continuous authentication model based on the touch behavior of the mobile terminal. The model uses the data collected by the accelerometer, gyroscope, and magnetometer to generate feature vectors and extracts the feature vectors containing macroscopic features, microscopic features, and joint features. And we design a convolutional bidirectional recurrent neural network model to distinguish the sensor feature vectors. On this basis, we perform various experiments on a large dataset Hand Movement, Orientation, and Grasp (HMOG) with different sensor characteristics. Compared with the most advanced models proposed recently, the results show that our model achieves an equal error rate (EER) of 0.53%, which significantly improves authentication accuracy.
大多数现有的身份验证技术都依赖于某些方式进行首次登录验证，例如个人识别号码(PIN)、轨迹或生物特征。然而，这些方式存在着大量的安全风险，使人们长期面临猜密码攻击、跟踪攻击和肩冲浪攻击。一旦非法用户伪造身份完成认证或绕过首次登录认证，其后续行为将变得不可控制。为了解决上述问题，我们提出了一种基于移动终端触摸行为的隐式连续认证模型。该模型利用加速度计、陀螺仪和磁力计采集的数据生成特征向量，提取包含宏观特征、微观特征和关节特征的特征向量。设计了一种卷积双向递归神经网络模型来区分传感器特征向量。在此基础上，我们对具有不同传感器特征的大型数据集Hand Movement, Orientation, and Grasp (HMOG)进行了各种实验。与目前最先进的模型相比，该模型的等误差率(EER)为0.53%，显著提高了认证精度。
{"title":"Implicit Continuous Authentication Model Based on Mobile Terminal Touch Behavior","authors":"Rui Mao, Heming Ji, Di Cheng, Xiaoyu Wang, Yan Wang, Degang Sun","doi":"10.1109/ISCC55528.2022.9913017","DOIUrl":"https://doi.org/10.1109/ISCC55528.2022.9913017","url":null,"abstract":"Most existing identity authentication technologies rely on some ways for the first login authentication, such as personal identification number (PIN), track, or biological characteristics. However, these ways exist plenty of security risks, which make people face password guessing attacks, trace attacks, and shoulder surfing attacks for a long time. Once the illegal users forge identity to complete authentication or bypass first login authentication, their subsequent behavior will become out of control. To solve the above problems, we propose an implicit continuous authentication model based on the touch behavior of the mobile terminal. The model uses the data collected by the accelerometer, gyroscope, and magnetometer to generate feature vectors and extracts the feature vectors containing macroscopic features, microscopic features, and joint features. And we design a convolutional bidirectional recurrent neural network model to distinguish the sensor feature vectors. On this basis, we perform various experiments on a large dataset Hand Movement, Orientation, and Grasp (HMOG) with different sensor characteristics. Compared with the most advanced models proposed recently, the results show that our model achieves an equal error rate (EER) of 0.53%, which significantly improves authentication accuracy.","PeriodicalId":309606,"journal":{"name":"2022 IEEE Symposium on Computers and Communications (ISCC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131506952","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 : 2022-06-30DOI: 10.1109/ISCC55528.2022.9913033
C. Comito
Social media has become a key tool for spreading the news, discussing ideas and comments on world events, playing a relevant role also in public health management, especially in epidemics surveillance like seasonal flu. Online social media actually can provide an important help in monitoring disease spreading as users self-report their health-related issues. Since the very first days of COVID-19 outbreak, people exchanged news, updates, sentiment and opinion about the pandemics. The paper describes a study aiming at evaluating the correlation of tweets with official COVID-19 data. Based on the outcomes of the correlation study, the paper proposes a forecasting model to predict the number of new daily COVID-19 cases. The approach is formulated as an autoregressive model that combines tweets and official COVID-19 data. A real-word dataset of tweets is used for the correlation study and to evaluate the performance of the forecasting model. Results shown the feasibility of the approach, highlighting the improvement obtained when tweets are integrated in the forecasting model, allowing to predict new COVID-19 cases in advance, on average 4–6 days before they were confirmed.
{"title":"Sensing Social Media to Forecast COVID-19 Cases","authors":"C. Comito","doi":"10.1109/ISCC55528.2022.9913033","DOIUrl":"https://doi.org/10.1109/ISCC55528.2022.9913033","url":null,"abstract":"Social media has become a key tool for spreading the news, discussing ideas and comments on world events, playing a relevant role also in public health management, especially in epidemics surveillance like seasonal flu. Online social media actually can provide an important help in monitoring disease spreading as users self-report their health-related issues. Since the very first days of COVID-19 outbreak, people exchanged news, updates, sentiment and opinion about the pandemics. The paper describes a study aiming at evaluating the correlation of tweets with official COVID-19 data. Based on the outcomes of the correlation study, the paper proposes a forecasting model to predict the number of new daily COVID-19 cases. The approach is formulated as an autoregressive model that combines tweets and official COVID-19 data. A real-word dataset of tweets is used for the correlation study and to evaluate the performance of the forecasting model. Results shown the feasibility of the approach, highlighting the improvement obtained when tweets are integrated in the forecasting model, allowing to predict new COVID-19 cases in advance, on average 4–6 days before they were confirmed.","PeriodicalId":309606,"journal":{"name":"2022 IEEE Symposium on Computers and Communications (ISCC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131842437","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 : 2022-06-30DOI: 10.1109/ISCC55528.2022.9912977
Yukun Zhou, Dezun Dong, Zhengbin Pang, Junhong Ye, Feng Jin
The widespread deployment of Remote Direct Memory Access (RDMA) in datacenter networks increases the stringency for convergence speed when congestion occurs. Fast convergence significantly reduces buffer occupancy, which in turn lessens the probability of triggering Priority-based Flow Control (PFC). Besides, the propagation delay becomes shorter with rapidly growing link speed, which correspondingly makes the queueing delay a major part of end-to-end latency. Fast convergence and low buffer occupancy become more essential for lowering queue delay and flow complete time. We present DQCC (Double-Q Congestion Control), a fast-converging congestion control scheme, which consists of two fundamental components: (i) an ECN-marking-ratio-based queue buffer occupancy estimating (QBOE) solution and (ii) a queue-building-rate driven rate adjustment (QDRA) mechanism to achieve fast convergence. We conduct extensive experiments to evaluate the performance of DQCC, and the results show that DQCC greatly accelerates the convergence process. DQCC achieves low tail latency and low buffer occupancy simultaneously.
RDMA (Remote Direct Memory Access)技术在数据中心网络中的广泛应用，提高了在发生拥塞时收敛速度的严密性。快速收敛显著减少了缓冲区占用，从而降低了触发基于优先级的流量控制(PFC)的概率。此外，随着链路速度的快速增长，传播延迟变得越来越短，相应地，排队延迟成为端到端延迟的主要部分。快速收敛和低缓冲区占用对于降低队列延迟和流完成时间至关重要。我们提出了一种快速收敛的拥塞控制方案DQCC (Double-Q拥塞控制)，它由两个基本组成部分组成:(i)基于ecn标记比率的队列缓冲区占用估计(QBOE)解决方案和(ii)实现快速收敛的队列构建速率驱动的速率调整(QDRA)机制。我们进行了大量的实验来评估DQCC的性能，结果表明DQCC大大加快了收敛过程。DQCC同时实现了低尾延迟和低缓冲区占用。
{"title":"Fast-Converging Congestion Control in Datacenter Networks","authors":"Yukun Zhou, Dezun Dong, Zhengbin Pang, Junhong Ye, Feng Jin","doi":"10.1109/ISCC55528.2022.9912977","DOIUrl":"https://doi.org/10.1109/ISCC55528.2022.9912977","url":null,"abstract":"The widespread deployment of Remote Direct Memory Access (RDMA) in datacenter networks increases the stringency for convergence speed when congestion occurs. Fast convergence significantly reduces buffer occupancy, which in turn lessens the probability of triggering Priority-based Flow Control (PFC). Besides, the propagation delay becomes shorter with rapidly growing link speed, which correspondingly makes the queueing delay a major part of end-to-end latency. Fast convergence and low buffer occupancy become more essential for lowering queue delay and flow complete time. We present DQCC (Double-Q Congestion Control), a fast-converging congestion control scheme, which consists of two fundamental components: (i) an ECN-marking-ratio-based queue buffer occupancy estimating (QBOE) solution and (ii) a queue-building-rate driven rate adjustment (QDRA) mechanism to achieve fast convergence. We conduct extensive experiments to evaluate the performance of DQCC, and the results show that DQCC greatly accelerates the convergence process. DQCC achieves low tail latency and low buffer occupancy simultaneously.","PeriodicalId":309606,"journal":{"name":"2022 IEEE Symposium on Computers and Communications (ISCC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129834661","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 : 2022-06-30DOI: 10.1109/ISCC55528.2022.9912983
Guojian Luo, J. Qu, Lina Zhang, Xiaoyu Fang, Yi Zhang, Tong Zhou
Increased needs for social security promote the development of video surveillance, appealing to the exigency of real-time detection of anomalous events. Considering the rarity and unpredictability of anomalous events, a classical strategy is to model normal data and detect outliers to the model. As a fundamental generative model for time series data, Hidden Markov models (HMM) have been employed in various fields such as speech recognition and video analysis. In this paper, we propose the use of Bayesian HMMs with Dirichlet mixtures which are arrayed along patched frames with Dirichlet distributions as emission probability functions. These spatially-aligned HMMs evolve in parallel, significantly reducing inference time. Learning algorithm based on Stochastic Variational Inference and Discrete Variable Enumeration is applied to our model for fast and robust inference. Experiments over the public UCSD dataset demonstrate the validity of this approach.
{"title":"Spatial-based Bayesian Hidden Markov Models with Dirichlet Mixtures for Video Anomaly Detection","authors":"Guojian Luo, J. Qu, Lina Zhang, Xiaoyu Fang, Yi Zhang, Tong Zhou","doi":"10.1109/ISCC55528.2022.9912983","DOIUrl":"https://doi.org/10.1109/ISCC55528.2022.9912983","url":null,"abstract":"Increased needs for social security promote the development of video surveillance, appealing to the exigency of real-time detection of anomalous events. Considering the rarity and unpredictability of anomalous events, a classical strategy is to model normal data and detect outliers to the model. As a fundamental generative model for time series data, Hidden Markov models (HMM) have been employed in various fields such as speech recognition and video analysis. In this paper, we propose the use of Bayesian HMMs with Dirichlet mixtures which are arrayed along patched frames with Dirichlet distributions as emission probability functions. These spatially-aligned HMMs evolve in parallel, significantly reducing inference time. Learning algorithm based on Stochastic Variational Inference and Discrete Variable Enumeration is applied to our model for fast and robust inference. Experiments over the public UCSD dataset demonstrate the validity of this approach.","PeriodicalId":309606,"journal":{"name":"2022 IEEE Symposium on Computers and Communications (ISCC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2022-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130878374","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：