Javier Bustos-Jiménez, Victor Ramiro, F. Lalanne, T. Barros
Broadband Internet access has become a regular commodity on households around the world. For this reason, initiatives for characterization and monitoring of service offerings by Internet Service Providers (ISPs), by governments, researchers and independent entities have grown in popularity in the last few years. In this context, we present Adkintun, the Chilean initiative commanded by the Transports and Telecommunications Ministry (MTT) of Chile to NIC Chile Research Labs, in order to characterize the growing broadband Internet offer in that country providing an independent tool for consumers to monitor the effective quality of their provided service, and perform informed choices among chilean internet providers. In this article, we present the design, architecture and main results of Adkintun after a year of monitoring, measuring and characterizing chilean internet.
{"title":"Adkintun: SLA Monitoring of ISP Broadband Offerings","authors":"Javier Bustos-Jiménez, Victor Ramiro, F. Lalanne, T. Barros","doi":"10.1109/WAINA.2013.240","DOIUrl":"https://doi.org/10.1109/WAINA.2013.240","url":null,"abstract":"Broadband Internet access has become a regular commodity on households around the world. For this reason, initiatives for characterization and monitoring of service offerings by Internet Service Providers (ISPs), by governments, researchers and independent entities have grown in popularity in the last few years. In this context, we present Adkintun, the Chilean initiative commanded by the Transports and Telecommunications Ministry (MTT) of Chile to NIC Chile Research Labs, in order to characterize the growing broadband Internet offer in that country providing an independent tool for consumers to monitor the effective quality of their provided service, and perform informed choices among chilean internet providers. In this article, we present the design, architecture and main results of Adkintun after a year of monitoring, measuring and characterizing chilean internet.","PeriodicalId":359251,"journal":{"name":"2013 27th International Conference on Advanced Information Networking and Applications Workshops","volume":"59 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125917563","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}
D. Lezzi, Roger Rafanell, E. Torres, R. Giovanni, I. Blanquer, Rosa M. Badia
In the last decades biology scientists have relied on their own resources and tools to run the experiments and store the results of the analysis. However, the explosion of big data and the growing availability of computational methods find an obstacle in the lack of computational and storage resources. Cloud computing platforms are emerging as potential solution to overcome these limitations, but adaptation of the applications to enable scientific users to benefit from resources acquired on demand is a complex process requiring multidisciplinary expertise. The EUBrazilOpenBio initiative is implementing an e-Infrastructure that provides biodiversity community with a rich set of computational and data resources exploiting existing cloud technologies from EU and Brazil. This paper presents the implementation of one of the two use cases selected, the environmental niche modeling by means of implementing such workflow through the COMPSs framework and its deployment on the EUBrazil OpenBio platform. The proposed approach has been evaluated on a Cloud test bed managed by the VENUS-C middleware.
{"title":"Programming Ecological Niche Modeling Workflows in the Cloud","authors":"D. Lezzi, Roger Rafanell, E. Torres, R. Giovanni, I. Blanquer, Rosa M. Badia","doi":"10.1109/WAINA.2013.6","DOIUrl":"https://doi.org/10.1109/WAINA.2013.6","url":null,"abstract":"In the last decades biology scientists have relied on their own resources and tools to run the experiments and store the results of the analysis. However, the explosion of big data and the growing availability of computational methods find an obstacle in the lack of computational and storage resources. Cloud computing platforms are emerging as potential solution to overcome these limitations, but adaptation of the applications to enable scientific users to benefit from resources acquired on demand is a complex process requiring multidisciplinary expertise. The EUBrazilOpenBio initiative is implementing an e-Infrastructure that provides biodiversity community with a rich set of computational and data resources exploiting existing cloud technologies from EU and Brazil. This paper presents the implementation of one of the two use cases selected, the environmental niche modeling by means of implementing such workflow through the COMPSs framework and its deployment on the EUBrazil OpenBio platform. The proposed approach has been evaluated on a Cloud test bed managed by the VENUS-C middleware.","PeriodicalId":359251,"journal":{"name":"2013 27th International Conference on Advanced Information Networking and Applications Workshops","volume":"54 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129833740","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}
We call "Violent" (from portuguese: velocity and direction), a context-aware routing protocol that aims at obtaining and processing environment data such as velocity and direction of nodes as they establish contacts with each other in delay-tolerant networks (DTNs). This context information is used by the proposed algorithm - along with the node's position and time of the last encounter with other nodes - to estimate the position of the recipients of messages. The methodology implemented by VeloSent enables to identify more precisely which of the current neighboring nodes is moving towards the destination node, thus improving the delivery rate. The results obtained from simulation using different mobility models show that VeloSent required a lower number of transmissions and message hops when compared to other well-known routing algorithms, thus making it a good candidate for use in environments with scarce resources.
{"title":"Routing Protocol Based on the Position, Velocity, and Direction of the Nodes","authors":"G. Andrade, L. Lima, A. Calsavara, C. Maziero","doi":"10.1109/WAINA.2013.138","DOIUrl":"https://doi.org/10.1109/WAINA.2013.138","url":null,"abstract":"We call \"Violent\" (from portuguese: velocity and direction), a context-aware routing protocol that aims at obtaining and processing environment data such as velocity and direction of nodes as they establish contacts with each other in delay-tolerant networks (DTNs). This context information is used by the proposed algorithm - along with the node's position and time of the last encounter with other nodes - to estimate the position of the recipients of messages. The methodology implemented by VeloSent enables to identify more precisely which of the current neighboring nodes is moving towards the destination node, thus improving the delivery rate. The results obtained from simulation using different mobility models show that VeloSent required a lower number of transmissions and message hops when compared to other well-known routing algorithms, thus making it a good candidate for use in environments with scarce resources.","PeriodicalId":359251,"journal":{"name":"2013 27th International Conference on Advanced Information Networking and Applications Workshops","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129265208","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}
T. Volkert, Manuel Osdoba, A. Mitschele-Thiel, M. Becke
In this paper, we propose a new approach to enhance today's hop-by-hop routing. The main goal of our work is to improve the efficiency of network usage and to support transmission requirements (TR) provided by applications. In the paper, we focus on non-functional requirements such as a desired upper bound for the end-to-end delay and a required data rate. Both are encoded as meta data in each data packet. Furthermore, we describe how these embedded (in-band) TRs can be used for QoS-oriented routing, implemented on each intermediate router. This directs packets along routes, providing only the network resources actually needed. In order to support these routing decisions based on TRs, we utilize data from an additional control plane, which provides QoS-related topology data from the network. For this purpose, we apply our idea of "Hierarchical Routing Management" (HRM). It implements a signaling of QoS-related route capabilities. In order to provide a scalable system, we apply hierarchical addressing and topology aggregation, resulting in fewer details in higher hierarchy levels. In detail, we describe how SCTP can be used to encode and transport the TRs for a multipath video-streaming scenario. Moreover, we explain the HRM internal signaling processes, which are required to create and use the HRM system. Compared to current IP routing, this allows for a better network efficiency and QoE at receiver side. Nevertheless, the HRM approach remains backward compatible to current application and router implementations. The paper shows that these benefits can be achieved at an acceptable overhead for packet processing.
{"title":"Multipath Video Streaming Based on Hierarchical Routing Management","authors":"T. Volkert, Manuel Osdoba, A. Mitschele-Thiel, M. Becke","doi":"10.1109/WAINA.2013.161","DOIUrl":"https://doi.org/10.1109/WAINA.2013.161","url":null,"abstract":"In this paper, we propose a new approach to enhance today's hop-by-hop routing. The main goal of our work is to improve the efficiency of network usage and to support transmission requirements (TR) provided by applications. In the paper, we focus on non-functional requirements such as a desired upper bound for the end-to-end delay and a required data rate. Both are encoded as meta data in each data packet. Furthermore, we describe how these embedded (in-band) TRs can be used for QoS-oriented routing, implemented on each intermediate router. This directs packets along routes, providing only the network resources actually needed. In order to support these routing decisions based on TRs, we utilize data from an additional control plane, which provides QoS-related topology data from the network. For this purpose, we apply our idea of \"Hierarchical Routing Management\" (HRM). It implements a signaling of QoS-related route capabilities. In order to provide a scalable system, we apply hierarchical addressing and topology aggregation, resulting in fewer details in higher hierarchy levels. In detail, we describe how SCTP can be used to encode and transport the TRs for a multipath video-streaming scenario. Moreover, we explain the HRM internal signaling processes, which are required to create and use the HRM system. Compared to current IP routing, this allows for a better network efficiency and QoE at receiver side. Nevertheless, the HRM approach remains backward compatible to current application and router implementations. The paper shows that these benefits can be achieved at an acceptable overhead for packet processing.","PeriodicalId":359251,"journal":{"name":"2013 27th International Conference on Advanced Information Networking and Applications Workshops","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129550804","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}
K. Asimakis, Christos Bouras, V. Kokkinos, Andreas Papazois
Cellular networks' Multimedia Broadcast/Multicast Service (MBMS) over a Single Frequency Network (MBSFN) allows broadcasting cell signals to constructively interfere and users to take advantage of increased bit rates compared to other broadcasting technologies. In this paper, we present our work on the optimization of the power control for future cellular networks that employ MBSFN transmission scheme. We propose a novel simulator and optimizer that can minimize the transmission power of individual cells taking into account the changing positions of users. In order to read descriptions of the various scenarios, the simulator uses the eXtendable Temporal Network Description Language (XTNDL), a language that is defined in this paper. The optimizer's engine can be used in conjunction with the simulator or can be modified to read real time measurements from a real network.
{"title":"Mobility-Sensitive Power Control for MBSFN Cellular Networks","authors":"K. Asimakis, Christos Bouras, V. Kokkinos, Andreas Papazois","doi":"10.1109/WAINA.2013.73","DOIUrl":"https://doi.org/10.1109/WAINA.2013.73","url":null,"abstract":"Cellular networks' Multimedia Broadcast/Multicast Service (MBMS) over a Single Frequency Network (MBSFN) allows broadcasting cell signals to constructively interfere and users to take advantage of increased bit rates compared to other broadcasting technologies. In this paper, we present our work on the optimization of the power control for future cellular networks that employ MBSFN transmission scheme. We propose a novel simulator and optimizer that can minimize the transmission power of individual cells taking into account the changing positions of users. In order to read descriptions of the various scenarios, the simulator uses the eXtendable Temporal Network Description Language (XTNDL), a language that is defined in this paper. The optimizer's engine can be used in conjunction with the simulator or can be modified to read real time measurements from a real network.","PeriodicalId":359251,"journal":{"name":"2013 27th International Conference on Advanced Information Networking and Applications Workshops","volume":"78 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131114787","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 last decade has seen an exponential rise in the use of wireless sensor networks (WSNs) in various applications. While these have been primarily used on their own, researchers are now looking into ways of integrating these WSNs with other existing communication technologies. One such network is the satellite network which provides significant advantage in providing communication access to remote locations due to their inherent large coverage areas. Combining WSNs and satellite will enable us to perform efficient remotely monitoring in areas where terrestrial networks may not be present. However in such a scenario, the placement of sensor nodes is crucial in order to ensure efficient routing and energy-efficiency. This paper presents four network architectures for sensor-satellite hybrid networks, sensor-satellite direct communication, connections via a gateway node employing random node layout, grid-based node layout and cluster-based node layout with data aggregation. These architectures were simulated using network simulator 2 (ns-2) and then their packet loss rate, average end-to-end packet delay, and overall energy consumption were compared. The paper concludes by proposing a suitable network topology for environmental monitoring applications.
{"title":"Performance Evaluation of Alternative Network Architectures for Sensor-Satellite Integrated Networks","authors":"Suraj Verma, P. Pillai, Yim-Fun Hu","doi":"10.1109/WAINA.2013.56","DOIUrl":"https://doi.org/10.1109/WAINA.2013.56","url":null,"abstract":"The last decade has seen an exponential rise in the use of wireless sensor networks (WSNs) in various applications. While these have been primarily used on their own, researchers are now looking into ways of integrating these WSNs with other existing communication technologies. One such network is the satellite network which provides significant advantage in providing communication access to remote locations due to their inherent large coverage areas. Combining WSNs and satellite will enable us to perform efficient remotely monitoring in areas where terrestrial networks may not be present. However in such a scenario, the placement of sensor nodes is crucial in order to ensure efficient routing and energy-efficiency. This paper presents four network architectures for sensor-satellite hybrid networks, sensor-satellite direct communication, connections via a gateway node employing random node layout, grid-based node layout and cluster-based node layout with data aggregation. These architectures were simulated using network simulator 2 (ns-2) and then their packet loss rate, average end-to-end packet delay, and overall energy consumption were compared. The paper concludes by proposing a suitable network topology for environmental monitoring applications.","PeriodicalId":359251,"journal":{"name":"2013 27th International Conference on Advanced Information Networking and Applications Workshops","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125640100","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}
Semantic Interoperability is one of the most important open challenges in the Knowledge Management field. In this work, we propose a framework for making easier information exchanging operations among heterogeneous information sources, over the Cloud. Exploiting the Semantic Web technologies, the framework has the objective of allowing semantic interoperability among distributed software agents preserving not only semantics of transmitted messages, but also the subjectivity of agent's world vision in the communication process. The proposed system, in fact, will help users in the digital documents exchanging activities by extracting the related relevant information and coding their informative contents in a shape capable of preserving the original semantics, also in the case of communications agents using different or incompatible vocabularies. The framework is a an evolution of semantic interoperable methodologies towards cloud computer technologies, thus guaranteeing several advantages in terms of usability, scalability and fault tolerance. In particular, we implemented a prototype that provides interoperability functionalities in the E-Health domain, managing information coming from heterogeneous sources.
{"title":"A Framework for Semantic Interoperability over the Cloud","authors":"F. Amato, A. Mazzeo, V. Moscato, A. Picariello","doi":"10.1109/WAINA.2013.218","DOIUrl":"https://doi.org/10.1109/WAINA.2013.218","url":null,"abstract":"Semantic Interoperability is one of the most important open challenges in the Knowledge Management field. In this work, we propose a framework for making easier information exchanging operations among heterogeneous information sources, over the Cloud. Exploiting the Semantic Web technologies, the framework has the objective of allowing semantic interoperability among distributed software agents preserving not only semantics of transmitted messages, but also the subjectivity of agent's world vision in the communication process. The proposed system, in fact, will help users in the digital documents exchanging activities by extracting the related relevant information and coding their informative contents in a shape capable of preserving the original semantics, also in the case of communications agents using different or incompatible vocabularies. The framework is a an evolution of semantic interoperable methodologies towards cloud computer technologies, thus guaranteeing several advantages in terms of usability, scalability and fault tolerance. In particular, we implemented a prototype that provides interoperability functionalities in the E-Health domain, managing information coming from heterogeneous sources.","PeriodicalId":359251,"journal":{"name":"2013 27th International Conference on Advanced Information Networking and Applications Workshops","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122141878","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}
Javier Reyes, J. Orduña, G. Vigueras, Rafael Tornero
The era of distributed computing, where applications are executed on platforms like clusters, grids and/or clouds of computers, have shown the need for taking into account the communications that take place on distributed computer architectures when executing applications. In that environment, different communication-aware mapping techniques were proposed for improving the system performance, both for off-chip and for on-chip networks. Some of these proposals are based on heuristic search for finding pseudo-optimal assignments of a given population of tasks and processing elements. The technology improvement has allowed a significant increase in the problem size, multiplying the number of processor cores in each chip. Therefore, the proposals based on heuristic search must be accelerated in order to search in larger exploration domains within the same execution times. In this paper, we propose a comparative study of the parallel version of the local search method for communication-aware task mapping techniques. Unlike other comparative studies of heuristic methods implemented on GPUs, we compare the performance provided by the parallel version for GPUs with the performance provided by a MPI parallel version in terms of execution times and fitness values provided. The MPI version was executed on a cluster optimized for MPI applications. Also, we have considered a GPU with Fermi architecture and we have mapped the local search algorithm onto the GPU in order to improve the performance. The results show that the parallel implementation on a single GPU provides similar fitness function values than the MPI implementation on the cluster. However, the execution times required by the GPU implementation are significantly lower than the ones required by the MPI implementation, and these differences increase as so does size of the parallel system.
{"title":"Acceleration of Communication-Aware Task Mapping Techniques through GPU Computing","authors":"Javier Reyes, J. Orduña, G. Vigueras, Rafael Tornero","doi":"10.1109/WAINA.2013.38","DOIUrl":"https://doi.org/10.1109/WAINA.2013.38","url":null,"abstract":"The era of distributed computing, where applications are executed on platforms like clusters, grids and/or clouds of computers, have shown the need for taking into account the communications that take place on distributed computer architectures when executing applications. In that environment, different communication-aware mapping techniques were proposed for improving the system performance, both for off-chip and for on-chip networks. Some of these proposals are based on heuristic search for finding pseudo-optimal assignments of a given population of tasks and processing elements. The technology improvement has allowed a significant increase in the problem size, multiplying the number of processor cores in each chip. Therefore, the proposals based on heuristic search must be accelerated in order to search in larger exploration domains within the same execution times. In this paper, we propose a comparative study of the parallel version of the local search method for communication-aware task mapping techniques. Unlike other comparative studies of heuristic methods implemented on GPUs, we compare the performance provided by the parallel version for GPUs with the performance provided by a MPI parallel version in terms of execution times and fitness values provided. The MPI version was executed on a cluster optimized for MPI applications. Also, we have considered a GPU with Fermi architecture and we have mapped the local search algorithm onto the GPU in order to improve the performance. The results show that the parallel implementation on a single GPU provides similar fitness function values than the MPI implementation on the cluster. However, the execution times required by the GPU implementation are significantly lower than the ones required by the MPI implementation, and these differences increase as so does size of the parallel system.","PeriodicalId":359251,"journal":{"name":"2013 27th International Conference on Advanced Information Networking and Applications Workshops","volume":"14 4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116432891","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}
Chiraz Houaidia, H. Idoudi, A. Bossche, T. Val, L. Saïdane
Current routing mechanisms proposed for adhocnetworks are still feasibly applied in Wireless Mesh Networks given their similarities. Many researchers have conducted numerous simulations for comparing the performances of these routing protocols under various conditions and constraints. Most made comparisons are not aware of PHY/MAC layers and their impact on routing performances. In this paper we study through simulations the impact of PHY/MAC protocols on higher layers. The considered protocols include three propagation models, i.e., Free Space, TwoRayGround and Shadowing, three different PHY/MAC protocols specified IEEE 802.11 standards namely, 802.11b, 802.11s and 802.11n, and finally three routing protocols, i.e., AODV, OLSR and HWMP. In a comparative way, we investigate the effectiveness of these protocols when they coexist on a wireless mesh network environment. Our results show that the routing strategy can significantly impact the network performance only if it is strongly linked to the characteristics of the lower layers.
{"title":"Impact of IEEE 802.11 PHY/MAC Strategies on Routing Performance in Wireless Mesh Networks","authors":"Chiraz Houaidia, H. Idoudi, A. Bossche, T. Val, L. Saïdane","doi":"10.1109/WAINA.2013.2","DOIUrl":"https://doi.org/10.1109/WAINA.2013.2","url":null,"abstract":"Current routing mechanisms proposed for adhocnetworks are still feasibly applied in Wireless Mesh Networks given their similarities. Many researchers have conducted numerous simulations for comparing the performances of these routing protocols under various conditions and constraints. Most made comparisons are not aware of PHY/MAC layers and their impact on routing performances. In this paper we study through simulations the impact of PHY/MAC protocols on higher layers. The considered protocols include three propagation models, i.e., Free Space, TwoRayGround and Shadowing, three different PHY/MAC protocols specified IEEE 802.11 standards namely, 802.11b, 802.11s and 802.11n, and finally three routing protocols, i.e., AODV, OLSR and HWMP. In a comparative way, we investigate the effectiveness of these protocols when they coexist on a wireless mesh network environment. Our results show that the routing strategy can significantly impact the network performance only if it is strongly linked to the characteristics of the lower layers.","PeriodicalId":359251,"journal":{"name":"2013 27th International Conference on Advanced Information Networking and Applications Workshops","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127806842","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}
Current application of GPU processors for parallel computing tasks show excellent results in terms of speed-ups compared to CPU processors. However, there is no existing framework that enables automatic distribution of data and processing across multiple GPUs, modularity of kernel design, and efficient co-usage of CPU and GPU processors. All these elements are necessary conditions to enable users to easily perform 'Big Data' analysis, and to create their own modules for their desired processing functionality. We propose a framework for in-memory 'Big Text Data' analytics that provides mechanisms for automatic data segmentation, distribution, execution, and result retrieval across multiple cards (CPU, GPU & FPGA) and machines, and a modular design for easy addition of new GPU kernels. The architecture and components of the framework such as multi-card data distribution and execution, data structures for efficient memory access, algorithms for parallel GPU computation, and result retrieval are described in detail, and some of the kernels in the framework are evaluated using Big Data versus multi-core CPUs to demonstrate the performance and feasibility of using it for 'Big Data' analytics, providing alternative and cheaper HPC solution.
{"title":"A Multi-GPU Framework for In-Memory Text Data Analytics","authors":"P. K. Chong, E. Karuppiah, K. Yong","doi":"10.1109/WAINA.2013.238","DOIUrl":"https://doi.org/10.1109/WAINA.2013.238","url":null,"abstract":"Current application of GPU processors for parallel computing tasks show excellent results in terms of speed-ups compared to CPU processors. However, there is no existing framework that enables automatic distribution of data and processing across multiple GPUs, modularity of kernel design, and efficient co-usage of CPU and GPU processors. All these elements are necessary conditions to enable users to easily perform 'Big Data' analysis, and to create their own modules for their desired processing functionality. We propose a framework for in-memory 'Big Text Data' analytics that provides mechanisms for automatic data segmentation, distribution, execution, and result retrieval across multiple cards (CPU, GPU & FPGA) and machines, and a modular design for easy addition of new GPU kernels. The architecture and components of the framework such as multi-card data distribution and execution, data structures for efficient memory access, algorithms for parallel GPU computation, and result retrieval are described in detail, and some of the kernels in the framework are evaluated using Big Data versus multi-core CPUs to demonstrate the performance and feasibility of using it for 'Big Data' analytics, providing alternative and cheaper HPC solution.","PeriodicalId":359251,"journal":{"name":"2013 27th International Conference on Advanced Information Networking and Applications Workshops","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127912398","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：