Pub Date : 2015-07-20DOI: 10.1109/HPCSim.2015.7237045
Xi Li, Anthony Ventresque, Jesús Omana Iglesias, John Murphy
Server consolidation is the most common and effective method to save energy and increase resource utilization in data centers, and virtual machine (VM) placement is the usual way of achieving server consolidation. VM placement is however challenging given the scale of IT infrastructures nowadays and the risk of resource contention among co-located VMs after consolidation. Therefore, the correlation among VMs to be co-located need to be considered. However, existing solutions do not address the scalability issue that arises once the number of VMs increases to an order of magnitude that makes it unrealistic to calculate the correlation between each pair of VMs. In this paper, we propose a correlation-aware VM consolidation solution ScalCCon1, which uses a novel two-phase clustering scheme to address the aforementioned scalability problem. We propose and demonstrate the benefits of using the two-phase clustering scheme in comparison to solutions using one-phase clustering (up to 84% reduction of execution time when 17, 446 VMs are considered). Moreover, our solution manages to reduce the number of physical machines (PMs) required, as well as the number of performance violations, compared to existing correlation-based approaches.
{"title":"Scalable correlation-aware virtual machine consolidation using two-phase clustering","authors":"Xi Li, Anthony Ventresque, Jesús Omana Iglesias, John Murphy","doi":"10.1109/HPCSim.2015.7237045","DOIUrl":"https://doi.org/10.1109/HPCSim.2015.7237045","url":null,"abstract":"Server consolidation is the most common and effective method to save energy and increase resource utilization in data centers, and virtual machine (VM) placement is the usual way of achieving server consolidation. VM placement is however challenging given the scale of IT infrastructures nowadays and the risk of resource contention among co-located VMs after consolidation. Therefore, the correlation among VMs to be co-located need to be considered. However, existing solutions do not address the scalability issue that arises once the number of VMs increases to an order of magnitude that makes it unrealistic to calculate the correlation between each pair of VMs. In this paper, we propose a correlation-aware VM consolidation solution ScalCCon1, which uses a novel two-phase clustering scheme to address the aforementioned scalability problem. We propose and demonstrate the benefits of using the two-phase clustering scheme in comparison to solutions using one-phase clustering (up to 84% reduction of execution time when 17, 446 VMs are considered). Moreover, our solution manages to reduce the number of physical machines (PMs) required, as well as the number of performance violations, compared to existing correlation-based approaches.","PeriodicalId":134009,"journal":{"name":"2015 International Conference on High Performance Computing & Simulation (HPCS)","volume":"17 3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-07-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126235310","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 : 2015-07-20DOI: 10.1109/HPCSim.2015.7237026
G. Stringhini
In this paper we highlighted some threats that affect the users of cloud-based online services. We then presented a handful of systems that we developed over the last years to detect and block malicious activity on online services. Although the threat landscape is constantly evolving, we believe that such techniques constitute a good foundation to increasing the security of online service users.
{"title":"On the threats to Cloud-based online service users (and what we can do about them)","authors":"G. Stringhini","doi":"10.1109/HPCSim.2015.7237026","DOIUrl":"https://doi.org/10.1109/HPCSim.2015.7237026","url":null,"abstract":"In this paper we highlighted some threats that affect the users of cloud-based online services. We then presented a handful of systems that we developed over the last years to detect and block malicious activity on online services. Although the threat landscape is constantly evolving, we believe that such techniques constitute a good foundation to increasing the security of online service users.","PeriodicalId":134009,"journal":{"name":"2015 International Conference on High Performance Computing & Simulation (HPCS)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-07-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129754296","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 : 2015-07-20DOI: 10.1109/HPCSim.2015.7237085
M. Mirto, Laura Conte, G. Aloisio, C. Distante, Pietro Vecchio, Alessandra De Giovanni
Phytoplankton is a quality element for determining the ecological status of transitional water ecosystems. In routine analysis, bio-volume and surface area of phytoplankton are the most studied morphometric descriptors. Bio-volume can be estimated by comparing the algae with similar three-dimensional geometric forms and determining their volume, by measuring the linear dimensions required for its calculation with images acquired by an inverse microscope. Software such as LUCIA-G (Laboratory Imaging) determines, in an automatic way, only the linear dimensions of simple forms such as circle or ellipse, approximated at a given algae, whereas complex forms require the intervention of an operator by selecting the start and end points of linear dimensions with obvious introduction of human error. In this paper, we propose a novel methodology for detecting phytoplankton algae and by measuring linear dimensions of 42 geometrical forms to automatically compute their area and bio-volume, that has been implemented in a novel software, named LUISA, for image analysis.
{"title":"Measuring cells in phytoplankton images","authors":"M. Mirto, Laura Conte, G. Aloisio, C. Distante, Pietro Vecchio, Alessandra De Giovanni","doi":"10.1109/HPCSim.2015.7237085","DOIUrl":"https://doi.org/10.1109/HPCSim.2015.7237085","url":null,"abstract":"Phytoplankton is a quality element for determining the ecological status of transitional water ecosystems. In routine analysis, bio-volume and surface area of phytoplankton are the most studied morphometric descriptors. Bio-volume can be estimated by comparing the algae with similar three-dimensional geometric forms and determining their volume, by measuring the linear dimensions required for its calculation with images acquired by an inverse microscope. Software such as LUCIA-G (Laboratory Imaging) determines, in an automatic way, only the linear dimensions of simple forms such as circle or ellipse, approximated at a given algae, whereas complex forms require the intervention of an operator by selecting the start and end points of linear dimensions with obvious introduction of human error. In this paper, we propose a novel methodology for detecting phytoplankton algae and by measuring linear dimensions of 42 geometrical forms to automatically compute their area and bio-volume, that has been implemented in a novel software, named LUISA, for image analysis.","PeriodicalId":134009,"journal":{"name":"2015 International Conference on High Performance Computing & Simulation (HPCS)","volume":"53 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-07-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133649745","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 : 2015-07-20DOI: 10.1109/HPCSim.2015.7237044
Syed Asif Raza Shah, A. Jaikar, S. Noh
Virtualization technology plays a vital role in cloud computing. One of the core features of virtualization technology is live virtual machine migration. The live migration is a process of transferring the complete state of virtual machine between physical hosts without any service interruption. This capability is being widely used for the purpose of system maintenance, load balancing, energy efficiency, reconfiguration and fault tolerance. Live migration has been extensively studied for commercial workloads and the ongoing research is mainly focusing on the performance improvements of well-known live migration algorithms. Today, scientific communities are actively thinking to take the advantage of cloud computing for scientific workloads. In this paper, we analyze the performance of well-known precopy, postcopy and hybrid live migration algorithms and examine the migration times of VMs running high throughput computing (HTC) jobs in a scientific cloud computing environment. The results of our research not only show the performance comparison of live migration algorithms but also will be helpful when selecting a live migration algorithm in scientific cloud computing environment.
{"title":"A performance analysis of precopy, postcopy and hybrid live VM migration algorithms in scientific cloud computing environment","authors":"Syed Asif Raza Shah, A. Jaikar, S. Noh","doi":"10.1109/HPCSim.2015.7237044","DOIUrl":"https://doi.org/10.1109/HPCSim.2015.7237044","url":null,"abstract":"Virtualization technology plays a vital role in cloud computing. One of the core features of virtualization technology is live virtual machine migration. The live migration is a process of transferring the complete state of virtual machine between physical hosts without any service interruption. This capability is being widely used for the purpose of system maintenance, load balancing, energy efficiency, reconfiguration and fault tolerance. Live migration has been extensively studied for commercial workloads and the ongoing research is mainly focusing on the performance improvements of well-known live migration algorithms. Today, scientific communities are actively thinking to take the advantage of cloud computing for scientific workloads. In this paper, we analyze the performance of well-known precopy, postcopy and hybrid live migration algorithms and examine the migration times of VMs running high throughput computing (HTC) jobs in a scientific cloud computing environment. The results of our research not only show the performance comparison of live migration algorithms but also will be helpful when selecting a live migration algorithm in scientific cloud computing environment.","PeriodicalId":134009,"journal":{"name":"2015 International Conference on High Performance Computing & Simulation (HPCS)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-07-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129464490","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 : 2015-07-20DOI: 10.1109/HPCSim.2015.7237087
Giovanni Simonini, Song Zhu
Big data analysis now drives nearly every aspect of modern society, from manufacturing and retail, through mobile and financial services, through the life sciences and physical sciences. The ability to continue to use big data to make new connections and discoveries will help to drive the breakthroughs of tomorrow. One of the most valuable means through which to make sense of big data, and thus make it more approachable to most people, is data visualization. Data visualization can guide decision-making and become a tool to convey information critical in all data analysis. However, to be actually actionable, data visualizations should contain the right amount of interactivity. They have to be well designed, easy to use, understandable, meaningful, and approachable. In this article we present a new approach to visualize huge amount of data, based on a Bayesian suggestion algorithm and the widely used enterprise search platform Solr. We demonstrate how the proposed Bayesian suggestion algorithm became a key ingredient in a big data scenario, where generally a query can generate so many results that the user can be confused. Thus, the selection of the best results, together with the result path chosen by the user by means of multi-faceted querying and faceted navigation, can be very useful.
{"title":"Big data exploration with faceted browsing","authors":"Giovanni Simonini, Song Zhu","doi":"10.1109/HPCSim.2015.7237087","DOIUrl":"https://doi.org/10.1109/HPCSim.2015.7237087","url":null,"abstract":"Big data analysis now drives nearly every aspect of modern society, from manufacturing and retail, through mobile and financial services, through the life sciences and physical sciences. The ability to continue to use big data to make new connections and discoveries will help to drive the breakthroughs of tomorrow. One of the most valuable means through which to make sense of big data, and thus make it more approachable to most people, is data visualization. Data visualization can guide decision-making and become a tool to convey information critical in all data analysis. However, to be actually actionable, data visualizations should contain the right amount of interactivity. They have to be well designed, easy to use, understandable, meaningful, and approachable. In this article we present a new approach to visualize huge amount of data, based on a Bayesian suggestion algorithm and the widely used enterprise search platform Solr. We demonstrate how the proposed Bayesian suggestion algorithm became a key ingredient in a big data scenario, where generally a query can generate so many results that the user can be confused. Thus, the selection of the best results, together with the result path chosen by the user by means of multi-faceted querying and faceted navigation, can be very useful.","PeriodicalId":134009,"journal":{"name":"2015 International Conference on High Performance Computing & Simulation (HPCS)","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-07-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133747644","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 : 2015-07-20DOI: 10.1109/HPCSim.2015.7237081
M. Chaieb, Jaber Jemai, K. Mellouli
We address the problem of modeling combinatorial optimization problems (COP). COPs are generally complex problems to solve. So a good modeling step is fundamental to make the solution easier. Our approach orients researches to choose the best modeling strategy from the beginning to avoid any problem in the solving process. This paper aims at proposing a new approach dealing with hard COPs particularly when the decomposition process leads to some well-known and canonical optimization sub-problems. We tried to draw a clear framework that will help to model hierarchical optimization problems. The framework will be composed by four decomposition strategies which are: objective based decomposition; constraints based decomposition, semantic decomposition and data partitioning strategy. For each strategy, we present supporting examples from the literature where it was applied. But, not all combinatorial problems can be benefit from the outcomes and benefits of modeling problems hierarchically, rather only particular problems can be modeled like a hierarchical optimization problem. Thus, we propose a set of decomposability conditions for decomposing COPs. Furthermore, we define the types of relationships between obtained sub-problems and how partial solutions can be merged to obtain the final solution.
{"title":"A four-decomposition strategies for hierarchically modeling combinatorial optimization problems: framework, conditions and relations","authors":"M. Chaieb, Jaber Jemai, K. Mellouli","doi":"10.1109/HPCSim.2015.7237081","DOIUrl":"https://doi.org/10.1109/HPCSim.2015.7237081","url":null,"abstract":"We address the problem of modeling combinatorial optimization problems (COP). COPs are generally complex problems to solve. So a good modeling step is fundamental to make the solution easier. Our approach orients researches to choose the best modeling strategy from the beginning to avoid any problem in the solving process. This paper aims at proposing a new approach dealing with hard COPs particularly when the decomposition process leads to some well-known and canonical optimization sub-problems. We tried to draw a clear framework that will help to model hierarchical optimization problems. The framework will be composed by four decomposition strategies which are: objective based decomposition; constraints based decomposition, semantic decomposition and data partitioning strategy. For each strategy, we present supporting examples from the literature where it was applied. But, not all combinatorial problems can be benefit from the outcomes and benefits of modeling problems hierarchically, rather only particular problems can be modeled like a hierarchical optimization problem. Thus, we propose a set of decomposability conditions for decomposing COPs. Furthermore, we define the types of relationships between obtained sub-problems and how partial solutions can be merged to obtain the final solution.","PeriodicalId":134009,"journal":{"name":"2015 International Conference on High Performance Computing & Simulation (HPCS)","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-07-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131911252","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 : 2015-07-20DOI: 10.1109/HPCSim.2015.7237062
Hugo Meyer, J. Sancho, W. Miao, H. Dorren, N. Calabretta, Montse Farreras
This paper analyzes the performance impact of Optical Packet Switches (OPS) on parallel HPC applications. Because these devices cannot store light, in case of a collision for accessing the same output port in the switch only one packet can proceed and the others are dropped. The analysis focuses on the negative impact of packet collisions in the OPS and subsequent re-transmissions of dropped packets. To carry out this analysis we have developed a system simulator that mimics the behavior of real HPC application traffic and optical network devices such as the OPS. By using real application traces we have analyzed how message re-transmissions could affect parallel executions. In addition, we have also developed a methodology that allows to process applications traces and determine packet concurrency. The concurrency evaluates the amount of simultaneous packets that applications could transmit in the network. Results have shown that there are applications that can benefit from the advantages of OPS technology. Taking into account the applications analyzed, these applications are the ones that show less than 1% of packet concurrency; whereas there are other applications where their performance could be impacted by up to 65%. This impact is mostly dependent on application traffic behavior that is successfully characterized by our proposed methodology.
{"title":"Performance evaluation of Optical Packet Switches on high performance applications","authors":"Hugo Meyer, J. Sancho, W. Miao, H. Dorren, N. Calabretta, Montse Farreras","doi":"10.1109/HPCSim.2015.7237062","DOIUrl":"https://doi.org/10.1109/HPCSim.2015.7237062","url":null,"abstract":"This paper analyzes the performance impact of Optical Packet Switches (OPS) on parallel HPC applications. Because these devices cannot store light, in case of a collision for accessing the same output port in the switch only one packet can proceed and the others are dropped. The analysis focuses on the negative impact of packet collisions in the OPS and subsequent re-transmissions of dropped packets. To carry out this analysis we have developed a system simulator that mimics the behavior of real HPC application traffic and optical network devices such as the OPS. By using real application traces we have analyzed how message re-transmissions could affect parallel executions. In addition, we have also developed a methodology that allows to process applications traces and determine packet concurrency. The concurrency evaluates the amount of simultaneous packets that applications could transmit in the network. Results have shown that there are applications that can benefit from the advantages of OPS technology. Taking into account the applications analyzed, these applications are the ones that show less than 1% of packet concurrency; whereas there are other applications where their performance could be impacted by up to 65%. This impact is mostly dependent on application traffic behavior that is successfully characterized by our proposed methodology.","PeriodicalId":134009,"journal":{"name":"2015 International Conference on High Performance Computing & Simulation (HPCS)","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-07-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128310225","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 : 2015-07-20DOI: 10.1109/HPCSim.2015.7237113
Bogdan Nicolae
Scientific and data-intensive computing have matured over the last couple of years in all fields of science and industry. Their rapid increase in complexity and scale has prompted ongoing efforts dedicated to reach exascale infrastructure capability by the end of the decade. However, advances in this context are not homogeneous: I/O capabilities in terms of networking and storage are lagging behind computational power and are often considered a major limitation that that persists even at petascale [1]. A particularly difficult challenge in this context are collective I/O access patterns (which we henceforth refer to as collective checkpointing) where all processes simultaneously dump large amounts of related data simultaneously to persistent storage. This pattern is often exhibited by large-scale, bulk-synchronous applications in a variety of circumstances, e.g., when they use checkpoint-restart fault tolerance techniques to save intermediate computational states at regular time intervals [2] or when intermediate, globally synchronized results are needed during the lifetime of the computation (e.g. to understand how a simulation progresses during key phases). Under such circumstances, a decoupled storage system (e.g. a parallel file system such as GPFS [3] or a specialized storage system such as BlobSeer [4]) does not provide sufficient I/O bandwidth to handle the explosion of data sizes: for example, Jones et al. [5] predict dump times in the order of several hours. In order to overcome the I/O bandwidth limitation, one potential solution is to equip the compute nodes with local storage (i.e., HDDs, SSDs, NVMs, etc.) or use I/O forwarding nodes. Using this approach, a large part of the data can be dumped locally, which completely avoids the need to consume and compete for the I/O bandwidth of a decoupled storage system. However, this is not without drawbacks: the local storage devices or I/O forwarding nodes are prone to failures and as such the data they hold is volatile. Thus, a popular approach in practice is to wait until the local dump has finished, then let the application continue while the checkpoints are in turn dumped to a parallel file system in background. Such a straightforward solution can be effective at hiding the overhead incurred to due I/O bandwidth limitations, but this not necessarily the case: it may happen that there is not enough time to fully flush everything to the parallel file system before the next collective checkpoint request is issued. In fact, this a likely scenario with growing scale, as the failure rate increases, which introduces the need to checkpoint at smaller intervals in order to compensate for this effect. Furthermore, a smaller checkpoint interval also means local dumps are frequent and as such their overhead becomes significant itself.
{"title":"Techniques to improve the scalability of collective checkpointing at large scale","authors":"Bogdan Nicolae","doi":"10.1109/HPCSim.2015.7237113","DOIUrl":"https://doi.org/10.1109/HPCSim.2015.7237113","url":null,"abstract":"Scientific and data-intensive computing have matured over the last couple of years in all fields of science and industry. Their rapid increase in complexity and scale has prompted ongoing efforts dedicated to reach exascale infrastructure capability by the end of the decade. However, advances in this context are not homogeneous: I/O capabilities in terms of networking and storage are lagging behind computational power and are often considered a major limitation that that persists even at petascale [1]. A particularly difficult challenge in this context are collective I/O access patterns (which we henceforth refer to as collective checkpointing) where all processes simultaneously dump large amounts of related data simultaneously to persistent storage. This pattern is often exhibited by large-scale, bulk-synchronous applications in a variety of circumstances, e.g., when they use checkpoint-restart fault tolerance techniques to save intermediate computational states at regular time intervals [2] or when intermediate, globally synchronized results are needed during the lifetime of the computation (e.g. to understand how a simulation progresses during key phases). Under such circumstances, a decoupled storage system (e.g. a parallel file system such as GPFS [3] or a specialized storage system such as BlobSeer [4]) does not provide sufficient I/O bandwidth to handle the explosion of data sizes: for example, Jones et al. [5] predict dump times in the order of several hours. In order to overcome the I/O bandwidth limitation, one potential solution is to equip the compute nodes with local storage (i.e., HDDs, SSDs, NVMs, etc.) or use I/O forwarding nodes. Using this approach, a large part of the data can be dumped locally, which completely avoids the need to consume and compete for the I/O bandwidth of a decoupled storage system. However, this is not without drawbacks: the local storage devices or I/O forwarding nodes are prone to failures and as such the data they hold is volatile. Thus, a popular approach in practice is to wait until the local dump has finished, then let the application continue while the checkpoints are in turn dumped to a parallel file system in background. Such a straightforward solution can be effective at hiding the overhead incurred to due I/O bandwidth limitations, but this not necessarily the case: it may happen that there is not enough time to fully flush everything to the parallel file system before the next collective checkpoint request is issued. In fact, this a likely scenario with growing scale, as the failure rate increases, which introduces the need to checkpoint at smaller intervals in order to compensate for this effect. Furthermore, a smaller checkpoint interval also means local dumps are frequent and as such their overhead becomes significant itself.","PeriodicalId":134009,"journal":{"name":"2015 International Conference on High Performance Computing & Simulation (HPCS)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-07-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122034306","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 : 2015-07-20DOI: 10.1109/HPCSim.2015.7237036
G. Mathey, P. Yébenes, P. García, F. Quiles, J. Escudero-Sahuquillo
Dragonfly topologies have gathered great interest as one of the most promising interconnection patterns for the networks at the core of HPC clusters. In this paper, we configure several simulated InfiniBand-based clusters with a Dragonfly topology, and we analyze the performance in these configurations of the routing engines included in the latest release of the InfiniBand Subnet Manager (OpenSM v3.3.19).
{"title":"Analyzing available routing engines for InfiniBand-based clusters with Dragonfly topology","authors":"G. Mathey, P. Yébenes, P. García, F. Quiles, J. Escudero-Sahuquillo","doi":"10.1109/HPCSim.2015.7237036","DOIUrl":"https://doi.org/10.1109/HPCSim.2015.7237036","url":null,"abstract":"Dragonfly topologies have gathered great interest as one of the most promising interconnection patterns for the networks at the core of HPC clusters. In this paper, we configure several simulated InfiniBand-based clusters with a Dragonfly topology, and we analyze the performance in these configurations of the routing engines included in the latest release of the InfiniBand Subnet Manager (OpenSM v3.3.19).","PeriodicalId":134009,"journal":{"name":"2015 International Conference on High Performance Computing & Simulation (HPCS)","volume":"89 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-07-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124751313","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 : 2015-07-20DOI: 10.1109/HPCSim.2015.7237118
M. Doering, L. Wolf
In many road traffic scenarios the ability to communicate among traffic participants is very helpful. Therefore, research and development in academia and industry in that field exists already for many years and is ongoing in several directions. Some examples are Vehicular Ad-hoc Networks (VANETs), e.g., using technologies like IEEE 802.11p, and vehicles communicating with backend systems, e.g., using 2/3/4G cellular networks. In opportunistic vehicular networks, vehicles may not only exchange data for the immediate use such as Cooperative Awareness Messages (CAMs) in the ETSI Intelligent Transport Systems (ITS). Instead, a more general type of network might be set up, also for application scenarios beyond direct road traffic related aspects. For instance, buses of public transportation systems could collect data from the field or distribute data among several buses. Thus, buses could become an important part of smart cities or Internet of Things (IoT) application scenarios. Important questions are then, e.g., how much data could be distributed in such a bus-based opportunistic network or how often is it possible to exchange data between buses. Usually, buses in urban public transport systems follow well planned but nevertheless highly dynamic schedules and trajectories. Thus, traffic conditions have a significant and complex influence on bus mobility, causing very characteristic movement properties that are considerably distinct from other road vehicles. Understanding these special characteristics is essential for the design and evaluation of opportunistic vehicular communication networks. For this purpose we inspect two large-scale bus movement traces and describe the available data and metadata. Moreover, we analyze and compare vehicle density, speed, update intervals, and characteristics that are specific to public transport. Especially for large cities, but even for smaller ones if many devices like vehicles, sensors, and various other IoT things are part of such a network, high-performance computing and simulation approaches are necessary to study, analyse, design, use and maintain such a system.
{"title":"Opportunistic vehicular networking: Large-scale bus movement traces as base for network analysis","authors":"M. Doering, L. Wolf","doi":"10.1109/HPCSim.2015.7237118","DOIUrl":"https://doi.org/10.1109/HPCSim.2015.7237118","url":null,"abstract":"In many road traffic scenarios the ability to communicate among traffic participants is very helpful. Therefore, research and development in academia and industry in that field exists already for many years and is ongoing in several directions. Some examples are Vehicular Ad-hoc Networks (VANETs), e.g., using technologies like IEEE 802.11p, and vehicles communicating with backend systems, e.g., using 2/3/4G cellular networks. In opportunistic vehicular networks, vehicles may not only exchange data for the immediate use such as Cooperative Awareness Messages (CAMs) in the ETSI Intelligent Transport Systems (ITS). Instead, a more general type of network might be set up, also for application scenarios beyond direct road traffic related aspects. For instance, buses of public transportation systems could collect data from the field or distribute data among several buses. Thus, buses could become an important part of smart cities or Internet of Things (IoT) application scenarios. Important questions are then, e.g., how much data could be distributed in such a bus-based opportunistic network or how often is it possible to exchange data between buses. Usually, buses in urban public transport systems follow well planned but nevertheless highly dynamic schedules and trajectories. Thus, traffic conditions have a significant and complex influence on bus mobility, causing very characteristic movement properties that are considerably distinct from other road vehicles. Understanding these special characteristics is essential for the design and evaluation of opportunistic vehicular communication networks. For this purpose we inspect two large-scale bus movement traces and describe the available data and metadata. Moreover, we analyze and compare vehicle density, speed, update intervals, and characteristics that are specific to public transport. Especially for large cities, but even for smaller ones if many devices like vehicles, sensors, and various other IoT things are part of such a network, high-performance computing and simulation approaches are necessary to study, analyse, design, use and maintain such a system.","PeriodicalId":134009,"journal":{"name":"2015 International Conference on High Performance Computing & Simulation (HPCS)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-07-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129099414","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：