Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE COMPUTER SOC  

Efficient utilization of multi-core processor architectures requires the development of parallel programs. However, achieving optimized parallelization remains challenging, often requiring specialized expertise and substantial development time. Various tools, including automatic parallelization techniques, aim to facilitate this process by analyzing source code and identifying parallelizable sections. Despite their utility, conventional tools relying on static analysis struggle to comprehensively handle all potential parallel patterns. In contrast, recent Large Language Models (LLMs) have demonstrated exceptional capabilities in Natural Language Processing (NLP) tasks, and have begun to be adapted to address software engineering problems. Leveraging this advancement, we propose an automatic parallelization approach for OpenMP framework using LLMs. Our method utilizes the CodeT5+ model, pre-trained on diverse programming languages, to analyze and parallelize C/C++ source code by generating OpenMP directives. To validate our approach, we trained the model on a dataset sourced from public GitHub repositories, focusing on implementations using OpenMP. Experimental evaluation focused on parallelizing benchmarks from the NASA Parallel Benchmark (NPB) suite demonstrates promising results: Among eight benchmarks evaluated, our approach achieved speedups comparable to manually parallelized codes in two cases, moderate speedups in three cases, and encountered a problem in effectively parallelizing one benchmark. Overall, our model successfully parallelized 73% of the loops in NPB, showing its potential to enhance productivity in parallel programming tasks.

DOI： 10.1109/CLUSTERWorkshops61563.2024.00028

Web of Science

Scopus

researchmap

Publishing type：Research paper (international conference proceedings)  

The exponential growth in networked systems, such as entertainment streaming services, has necessitated advancements in network performance. Traditional singlepath routing often results in bottlenecks and inefficiencies, whereas multipath routing can significantly enhance bandwidth utilization and accommodate diverse user demands. This study explores the integration of SRv6 (Segment Routing over IPv6) with MPTCP (Multipath TCP) to leverage multiple network routes concurrently. SRv6 offers a scalable routing architecture by embedding routing information within IPv6 headers, while MPTCP distributes traffic across multiple paths, enhancing both network performance and reliability. We conducted experiments using two network topologies: a synthetic topology with 7 nodes and 11 links, and the Internet2 OS3E topology with 34 nodes and 42 links. Throughput measurements were taken under four scenarios: single-path TCP without SRv6, single-path TCP with SRv6, MPTCP without SRv6, and MPTCP with SRv6. The experimental results confirmed that the integration of SRv6 with MPTCP achieved the expected improvements in network performance, demonstrating enhanced throughput and better resource utilization compared to single-path TCP configurations. The implications of this research indicate that SRv6 and MPTCP integration enhances throughput, network reliability, and efficiency. This approach provides additional perspective for further scalable solutions that can be applied across various network topologies, and addressing modern network demands. However, this research also identifies potential challenges such as throughput fluctuations when SRv6 and MPTCP are combined, which may affect the performance of realtime applications. Future research should explore these challenges to optimize the integration of SRv6 and MPTCP.

DOI： 10.1109/COMNETSAT63286.2024.10862254

Scopus

researchmap

Publishing type：Research paper (international conference proceedings)   Publisher：Springer Nature Singapore  

Parallel programming is essential to utilize multi-core processors but remains challenging because it requires extensive knowledge of both software and hardware. Various automatic parallelization tools based on static analysis have been developed to ease the development of parallel programs. However, hand-parallelized codes still outperform auto-parallelized codes. Meanwhile, transformer-based large language models have made ground-breaking progress in coder understanding and generation tasks. In this paper, we fine-tune a transformer-based code understanding model, CodeT5, to create a model for automatically identifying parallelizable for-loops. The trained model helps developers to identify independent for-loops that can be potentially parallelized using tools such as OpenMP to improve the program performance. Our model is trained over 90,908 for-loops collected from 9 million C/C++ source files of public GitHub repositories, and achieves a 0.895 F1 score in identifying parallelizable for-loops in public GitHub projects and a 0.713 F1 score in the NAS Parallel Benchmark suite.

DOI： 10.1007/978-981-99-8211-0_4

Scopus

researchmap

Publishing type：Research paper (international conference proceedings)   Publisher：Springer Nature Switzerland  

In high-performance computing (HPC), using Linux containers can improve the convenience of system management and utilization, as well as facilitate the adoption of HPC DevOps and Clouds. However, utilizing containers with MPI applications poses some challenges in building portable MPI containers that can leverage various high-performance interconnect hardware available on different systems. A common solution is to replace the entire MPI stack in the container, but this may introduce compatibility issues between the host and container libraries. An alternative and more flexible solution is to inject host-shared libraries into the container and use them in conjunction with the container’s MPI stack, commonly referred to as a hybrid approach. In this paper, we propose a hybrid solution that uses the UCX communication framework as an intermediary between MPI and various high-end interconnects. This solution simplifies the container-building process while still achieving near-native communication performance. The flexibility of the proposed method was tested by performing a cross-version injection experiment. The results showed a good degree of forward/backward compatibility between versions, which can also refer to good portability as well. We evaluated the performance of our solution using communication micro-benchmarks and application benchmarks, and found that it achieved near-native communication performance. We also compared our solution with a similar solution using the Libfabric communication library, and confirmed that they have similar pros and cons with similar performance characteristics in most cases.

DOI： 10.1007/978-3-031-46970-1_21

Scopus

researchmap

Publishing type：Research paper (scientific journal)   Publisher：Springer Science and Business Media LLC  

Virtual machines (VMs) are used extensively in the cloud. The underlying hypervisors allow hardware resources to be split into multiple virtual units which enables server consolidation, fault containment, and resource management. However, VMs with traditional architecture introduce heavy overhead and reduce application performance. Containers are becoming popular options for running applications, yet such a solution raises security concerns due to weaker isolation than VMs. We are at the point of container and traditional virtualization convergence where lightweight hypervisors are implemented and integrated into the container ecosystem to maximize the benefits of VM isolation and container performance. However, there has been no comprehensive comparison among different convergence architectures. To identify limitations and best-fit use cases, we investigate the characteristics of Docker, Kata, gVisor, Firecracker, and QEMU/KVM by measuring the performance of disk storage, main memory, CPU, network, system call, and startup time. In addition, we evaluate their performance of running the Nginx web server and the MySQL database management system. We use QEMU/KVM as an example of running traditional VMs, Docker as the standard runc container, and the rest as the representatives of lightweight hypervisors. We compare and analyze the benchmark results, discuss the possible implications, explain the trade-off each organization made, and elaborate on the pros and cons of each architecture.

DOI： 10.1007/s42979-023-01827-9

Scopus

researchmap

Other Link： https://link.springer.com/article/10.1007/s42979-023-01827-9/fulltext.html

Publishing type：Research paper (international conference proceedings)  

In distributed Software Defined Networking (SDN), multiple controllers need to maintain a consistent view of the network state among the controllers using consensus algorithms, which introduces additional communication overhead and network latency, especially in large-scale networks. Therefore, it is difficult to optimize the placement of controllers considering not only the delay between switches and controllers, but also the delay of consensus algorithms. This research utilizes Open Network Operating System (ONOS) as a platform for building distributed SDN, which employs an Atomix cluster to store and share the network state views across multiple controllers using a consensus algorithm. Therefore, the optimal placement problem in ONOS involves both the placement of ONOS controllers and the placement of Atomix nodes, which makes the problem more complex. In this research, we develop a practical model to estimate the Flow Setup Time (FST) as a measure of ONOS controller response time, and propose a method to determine the optimal placements of controllers and Atomix nodes by a multi-objective optimization. Moreover, we apply our method to several real-world networks on the Internet and visualize the experimental results. Based on the observed variations in FST across different network topologies, we provide a suggestion to improve ONOS performance.

DOI： 10.1109/CANDAR60563.2023.00017

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/candar/candar2023.html#KangTNII23

J-GLOBAL

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

Empirical Dynamic Modeling (EDM) is a mathematical framework for modeling and predicting non-linear time series data. Although EDM is increasingly adopted in various research fields, its application to large-scale data has been limited due to its high computational cost. This article presents kEDM, a high-performance implementation of EDM for analyzing large-scale time series datasets. kEDM adopts the Kokkos performance-portable programming model to efficiently run on both CPU and GPU while sharing a single code base. We also conduct hardware-specific optimization of performance-critical kernels. kEDM achieved up to $6.58\times $ speedup in pairwise causal inference of real-world biology datasets compared to an existing EDM implementation. Furthermore, we integrate multiple approximate k-NN search algorithms into EDM to enable the analysis of extremely large datasets that were intractable with conventional EDM based on exhaustive k-NN search. EDM-based time series forecast enhanced with approximate k-NN search demonstrated up to $790\times $ speedup compared to conventional Simplex projection with less than 1% increase in MAPE.

DOI： 10.1109/ACCESS.2023.3289836

Scopus

researchmap

Language：English   Publishing type：Research paper (scientific journal)   Publisher：The Institute of Electronics, Information and Communication Engineers  

Solar power is the most widely used renewable energy source, which reduces pollution consequences from using conventional fossil fuels. However, supplying stable power from solar power generation remains challenging because it is difficult to forecast power generation. Accurate prediction of solar power generation would allow effective control of the amount of electricity stored in batteries, leading in a stable supply of electricity. Although the number of power plants is increasing, building a solar power prediction model for a newly constructed power plant usually requires collecting a new training dataset for the new power plant, which takes time to collect a sufficient amount of data. This paper aims to develop a highly accurate solar power prediction model for multiple power plants available for both new and existing power plants. The proposed method trains the model on existing multiple power plants to generate a general prediction model, and then uses it for a new power plant while waiting for the data to be collected. In addition, the proposed method tunes the general prediction model on the newly collected dataset and improves the accuracy for the new power plant. We evaluated the proposed method on 55 power plants in Japan with the dataset collected for two and a half years. As a result, the pre-trained models of our proposed method significantly reduces the average RMSE of the baseline method by 73.19%. This indicates that the model can generalize over multiple power plants, and training using datasets from other power plants is effective in reducing the RMSE. Fine-tuning the pre-trained model further reduces the RMSE by 8.12%.

DOI： 10.1587/transcom.2022EBT0003

Scopus

researchmap

Publishing type：Research paper (international conference proceedings)  

Infrastructure as Code (IaC) has become the de-facto standard method for managing cloud resources. Just like general source code (e.g., Java, etc.), infrastructure code also has numerous bugs so it needs to be tested. While several testing frameworks for IaC for cloud systems have been developed in practice, researchers have paid little attention to their testing. This study presents an empirical investigation of the use of tests for IaC for cloud systems. Our empirical results show that (i) 55.2% of the repositories using Terratest have at least one server infrastructure test; (ii) developers often maintain server infrastructure tests (1.7%-11.3% commits out of all the commits); (iii) many repositories have tests for system functionality (28%), deployment (20%), and configuration (17%).

DOI： 10.1109/APSEC60848.2023.00075

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/apsec/apsec2023.html#SuwanachotePKILRMI23

Language：English   Publishing type：Research paper (scientific journal)   Publisher：The Institute of Electronics, Information and Communication Engineers  

OpenFlow is a widely adopted implementation of the Software-Defined Networking (SDN) architecture. Since conventional network monitoring systems are unable to cope with OpenFlow networks, researchers have developed various monitoring systems tailored for Open- Flow networks. However, these existing systems either rely on a specific controller framework or an API, both of which are not part of the Open- Flow specification, and thus limit their applicability. This article proposes a transparent and low-overhead monitoring system for OpenFlow networks, referred to as Opimon. Opimon monitors the network topology, switch statistics, and flow tables in an OpenFlow network and visualizes the result through a web interface in real-time. Opimon monitors a network by interposing a proxy between the controller and switches and intercepting every OpenFlow message exchanged. This design allows Opimon to be compatible with any OpenFlow switch or controller. We tested the functionalities of Opimon on a virtual network built using Mininet and a large-scale international OpenFlow testbed (PRAGMA-ENT). Furthermore, we measured the performance overhead incurred by Opimon and demonstrated that the overhead in terms of latency and throughput was less than 3% and 5%, respectively.

DOI： 10.1587/transcom.2021EBP3083

Scopus

researchmap

Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

The optimal path varies depending on the communication characteristics of each application. However, existing routing protocols such as BGP and OSPF do not take this fact into account. Although Software Defined Networking (SDN) has been considered as a possible solution to this problem, SDN technologies relying on a centralized controller has scalability issues. SRv6, which is a source routing protocol that enables SDN, handles routing decisions in a decentralized manner and is expected to scale better than previous SDN technologies.This paper proposes Acar, an adaptive routing system using SRv6 that adaptively controls routing by considering the bandwidth requirements of applications and the link utilization of the network. We conducted experiments on a virtual network and demonstrated that Acar achieves better load balancing between links and higher throughput compared to ECMP.

DOI： 10.1109/CCNC49033.2022.9700608

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/ccnc/ccnc2022.html#SugiuraTII22

J-GLOBAL

researchmap

Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

Federated learning trains a model on a centralized server using datasets distributed over a massive amount of edge devices. Since federated learning does not send local data from edge devices to the server, it preserves data privacy. It transfers the local models from edge devices instead of the local data. However, communication costs are frequently a problem in federated learning. This paper proposes a novel method to reduce the required communication cost for federated learning by transferring only top updated parameters in neural network models. The proposed method allows adjusting the criteria of updated parameters to trade-off the reduction of communication costs and the loss of model accuracy. We evaluated the proposed method using diverse models and datasets and found that it can achieve comparable performance to transfer original models for federated learning. As a result, the proposed method has achieved a reduction of the required communication costs around 90% when compared to the conventional method for VGG16. Furthermore, we found out that the proposed method is able to reduce the communication cost of a large model more than of a small model due to the different threshold of updated parameters in each model architecture.

DOI： 10.1109/ICFEC54809.2022.00008

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/icfec/icfec2022.html#ThonglekTINLI22

Publishing type：Research paper (scientific journal)  

Deploying neural network models to edge devices is becoming increasingly popular because such deployment decreases the response time and ensures better data privacy of services. However, running large models on edge devices poses challenges because of limited computing resources and storage space. Researchers have therefore proposed various model compression methods to reduce the model size. To balance the trade-off between model size and accuracy, conventional model compression methods require manual effort to find the optimal configuration that reduces the model size without significant degradation of accuracy. In this article, we propose a method to automatically find the optimal configurations for quantization. The proposed method suggests multiple compression configurations that produce models with different size and accuracy, from which users can select the configurations that suit their use cases. Additionally, we propose a retraining method that does not require any labeled datasets for retraining. We evaluated the proposed method using various neural network models for classification, regression and semantic similarity tasks, and demonstrated that the proposed method reduced the size of models by at least 30% while maintaining less than 1% loss of accuracy. We compared the proposed method with state-of-the-art automated compression methods, and showed that it can provide better compression configurations than existing methods.

DOI： 10.1109/ACCESS.2022.3190627

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：ACM  

Empirical Dynamic Modeling (EDM) is a state-of-the-art non-linear time-series analysis framework. Despite its wide applicability, EDM was not scalable to large datasets due to its expensive computational cost. To overcome this obstacle, researchers have attempted and succeeded in accelerating EDM from both algorithmic and implementational aspects. In previous work, we developed a massively parallel implementation of EDM targeting HPC systems (mpEDM). However, mpEDM maintains different backends for different architectures. This design becomes a burden in the increasingly diversifying HPC systems, when porting to new hardware. In this paper, we design and develop a performance-portable implementation of EDM based on the Kokkos performance portability framework (kEDM), which runs on both CPUs and GPUs while based on a single codebase. Furthermore, we optimize individual kernels specifically for EDM computation, and use real-world datasets to demonstrate up to 5.5 × speedup compared to mpEDM in convergent cross mapping computation.

DOI： 10.1145/3437359.3465571

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/xsede/pearc2021.html#TakahashiWIPTHS21

Language：English  

researchmap

Other Link： https://ndlsearch.ndl.go.jp/books/R000000004-I031546668

Language：Japanese   Publishing type：Research paper (scientific journal)   Publisher：Japan Society for Software Science and Technology  

In PBL, it is often pointed out that it is difficult to qualitatively evaluate the improvement of students' fundamental competencies for working persons. In this study, we propose a method to evaluate students' growth and to improve classes through self-evaluation by students and objective evaluation by a third party as an approach to qualitative evaluation of PBL courses. In the proposed method, we aim to evaluate the validity of the subjective self-evaluation by the students before and after taking the PBL course, based on the objective evaluation. In this paper, we report on the evaluation results of the PBL course in 2017 and 2018 by the proposed method using the self-evaluation of fundamental competencies for working persons before and after taking the PBL course and the objective evaluation through PROG competency test.

DOI： 10.11309/jssst.38.1_52

Scopus

CiNii Books

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper (scientific journal)   Publisher：Japan Society for Software Science and Technology  

<p>In PBL, it is often pointed out that it is difficult to qualitatively evaluate the improvement of students' fundamental competencies for working persons. In this study, we propose a method to evaluate students' growth and to improve classes through self-evaluation by students and objective evaluation by a third party as an approach to qualitative evaluation of PBL courses. In the proposed method, we aim to evaluate the validity of the subjective self-evaluation by the students before and after taking the PBL course, based on the objective evaluation. In this paper, we report on the evaluation results of the PBL course in 2017 and 2018 by the proposed method using the self-evaluation of fundamental competencies for working persons before and after taking the PBL course and the objective evaluation through PROG competency test.</p>

DOI： 10.11309/jssst.38.1_52

Scopus

CiNii Books

researchmap

Publishing type：Research paper (scientific journal)   Publisher：Studies in Computational Intelligence  

Real-time processing is a fast and prompt processing technology that needs to complete the execution within a limited time constraint almost equal to the input time. Executing such real-time processing needs an efficient auto-scaling system which provides sufficient resources to compute the process within the time constraint. We use Apache Spark framework to build a cluster which supports real-time processing. The major challenge of scaling Apache Spark cluster automatically for the real-time processing is how to handle the unpredictable input data size and also the unpredictable resource availability of the underlying cloud infrastructure. If the scaling-out of the cluster is too slow then the application can not be executed within the time constraint as a result of insufficient resources. If the scaling-in of the cluster is slow, the resources are wasted without being utilized, and it leads less resource utilization. This research follows the real-world scenario where the computing resources are bounded by a certain number of computing nodes due to limited budget as well as the computing time is limited due to the nature of near real-time application. We design an auto-scaling system that applies a deep reinforcement learning technique, DQN (Deep Q-Network), to improve resource utilization efficiently. Our model-based DQN allows to automatically optimize the scaling of the cluster, because the DQN can autonomously learn the given environment features so that it can take suitable actions to get the maximum reward under the limited execution time and worker nodes.

DOI： 10.1007/978-3-030-58930-1_23

Scopus

researchmap

J-GLOBAL

researchmap

Publishing type：Research paper (international conference proceedings)  

Solar power is the most widely used green energy. However, using solar power generation as a stable power supply remains challenging since the power output is difficult to predict. Accurate prediction of solar power generation enables efficient control of the amount of stored electricity in batteries to produce a stable supply of electricity. This paper aims to build a highly accurate solar power prediction model. For this purpose, we design a neural network model based on Long Short-Term Memory (LSTM) to predict the future solar power generation using past solar power generation and weather forecasts. Since a large and diverse dataset is required to train an accurate prediction model, we develop a neural network based on Generative Adversarial Network (GAN) to generate artificial datasets from the original training dataset to increase the amount and diversity of the training dataset. Additionally, stratified k-fold cross-validation is used to eliminate learning deviation during training. As a result, the proposed neural network model based on GAN improved the R2 score of LSTM from 0.750 to 0.805 with stratified k-fold cross-validation.

DOI： 10.1109/IGESSC53124.2021.9618702

Scopus

researchmap

Publishing type：Research paper (international conference proceedings)   Publisher：SCITEPRESS - Science and Technology Publications  

Virtual Machines (VMs) are used extensively in cloud computing. The underlying hypervisor allows hardware resources to be split into multiple virtual units which enhances resource utilization. However, VMs with traditional architecture introduce heavy overhead and reduce application performance. Containers have been introduced to overcome this drawback, yet such a solution raises security concerns due to poor isolation. Lightweight hypervisors have been leveraged to strike a balance between performance and isolation. However, there has been no comprehensive performance comparison among them. To identify the best fit use case, we investigate the performance characteristics of Docker container, Kata containers, gVisor, Firecracker and QEMU/KVM by measuring the performance on disk storage, main memory, CPU, network, system call and startup time. In addition, we evaluate their performance of running Nginx web server and MySQL database management system. We use QEMU/KVM as an example of traditional VM, Docker as the standard container and the rest as the representatives of lightweight hypervisors. We compare and analyze the benchmarking results, discuss the possible implications, explain the trade-off each organization made and elaborate on the pros and cons of each architecture.

DOI： 10.5220/0010440502150223

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/closer/closer2021.html#LiTIITP21

Publishing type：Research paper (international conference proceedings)   Publisher：SCITEPRESS  

Virtual Machines (VMs) are used extensively in cloud computing. The underlying hypervisor allows hardware resources to be split into multiple virtual units which enhances resource utilization. However, VMs with traditional architecture introduce heavy overhead and reduce application performance. Containers have been introduced to overcome this drawback, yet such a solution raises security concerns due to poor isolation. Lightweight hypervisors have been leveraged to strike a balance between performance and isolation. However, there has been no comprehensive performance comparison among them. To identify the best fit use case, we investigate the performance characteristics of Docker container, Kata containers, gVisor, Firecracker and QEMU/KVM by measuring the performance on disk storage, main memory, CPU, network, system call and startup time. In addition, we evaluate their performance of running Nginx web server and MySQL database management system. We use QEMU/KVM as an example of traditional VM, Docker as the standard container and the rest as the representatives of lightweight hypervisors. We compare and analyze the benchmarking results, discuss the possible implications, explain the trade-off each organization made and elaborate on the pros and cons of each architecture.

DOI： 10.5220/0010440502150223

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/closer/closer2021.html#LiTIITP21

Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

Empirical Dynamic Modeling (EDM) is a nonlinear time series causal inference framework. The latest implementation of EDM, cppEDM, has only been used for small datasets due to computational cost. With the growth of data collection capabilities, there is a great need to identify causal relationships in large datasets. We present mpEDM, a parallel distributed implementation of EDM optimized for modern GPU-centric supercomputers. We improve the original algorithm to reduce redundant computation and optimize the implementation to fully utilize hardware resources such as GPUs and SIMD units. As a use case, we run mpEDM on AI Bridging Cloud Infrastructure (ABCI) using datasets of an entire animal brain sampled at single neuron resolution to identify dynamical causation patterns across the brain. mpEDM is 1, 530× faster than cppEDM and a dataset containing 101, 729 neuron was analyzed in 199 seconds on 512 nodes. This is the largest EDM causal inference achieved to date.

DOI： 10.1109/ICPADS51040.2020.00035

Scopus

arXiv

researchmap

Other Link： https://dblp.uni-trier.de/db/journals/corr/corr2011.html#abs-2011-11082

Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

Federated learning trains a model on a centralized server using datasets distributed over a large number of edge devices. Applying federated learning ensures data privacy because it does not transfer local data from edge devices to the server. Existing federated learning algorithms assume that all deployed models share the same structure. However, it is often infeasible to distribute the same model to every edge device because of hardware limitations such as computing performance and storage space. This paper proposes a novel federated learning algorithm to aggregate information from multiple heterogeneous models. The proposed method uses weighted average ensemble to combine the outputs from each model. The weight for the ensemble is optimized using black box optimization methods. We evaluated the proposed method using diverse models and datasets and found that it can achieve comparable performance to conventional training using centralized datasets. Furthermore, we compared six different optimization methods to tune the weights for the weighted average ensemble and found that tree parzen estimator achieves the highest accuracy among the alternatives.

DOI： 10.1109/ICMLA51294.2020.00120

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/icmla/icmla2020.html#ThonglekTIIN20

Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

Running neural network models on edge devices is attracting much attention by neural network researchers since edge computing technology is becoming more powerful than ever. However, deploying large neural network models on edge devices is challenging due to the limitation in available computing resources and storage space. Therefore, model compression techniques have been recently studied to reduce the model size and fit models on resource-limited edge devices. Compressing neural network models reduces the size of a model, but also degrades the accuracy of the model since it reduces the precision of weights in the model. Consequently, a retraining method is required to recover the accuracy of compressed models. Most existing retraining methods require the original labeled training datasets to retrain the models, but labeling is a time-consuming process. In particular, we cannot always access the original labeled datasets because of privacy policies and license limitations. In this paper, we propose a method to retrain a compressed neural network model with an unlabeled dataset that is different from the original labeled dataset. We compress the neural network model using quantization to decrease the size of the model. Subsequently, the compressed model is retrained by our proposed retraining method without using a labeled dataset to recover the accuracy of the model. We compared the proposed retraining method against the conventional retraining. The proposed method reduced the size of VGG-16 and ResNet-50 by 81.10% and 52.45%, respectively without significant accuracy loss. In addition, our proposed retraining method is clearly faster than the conventional retraining method.

DOI： 10.1109/IJCNN48605.2020.9207190

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/ijcnn/ijcnn2020.html#ThonglekTINNTI20

Language：Japanese   Publishing type：Research paper (scientific journal)   Publisher：Japan Society for Software Science and Technology  

<p>AiBiC Kansai offers an education course, AiBiC Spiral, designed to develop abilities to solve individual social issues using Big Data, artificial intelligence, and Cloud technologies through teamwork. In this paper, we report on the problems of the environments for the education program and the discrepancies in the expectations on topics and contents found by the analysis of the questionnaire conducted in 2017, and describe the activities and results performed to improve the program in 2018. Through the improvements, we have found that the problem on the program environments has been dissolved. However, as for the discrepancies in the expectations, we have revealed that only some part of the problem was improved and there were still some students found the topics and contents provided by the program were different from their expectations.</p>

DOI： 10.11309/jssst.37.1_19

Scopus

CiNii Books

J-GLOBAL

researchmap

Language：Japanese  

J-GLOBAL

researchmap

Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

Empirical Dynamic Modeling (EDM) is a nonlinear time series causal inference framework. The latest implementation of EDM, cppEDM, has only been used for small datasets due to computational cost. With the growth of data collection capabilities, there is a great need to identify causal relationships in large datasets. We present mpEDM, a parallel distributed implementation of EDM optimized for modern GPU-centric supercomputers. We improve the original algorithm to reduce redundant computation and optimize the implementation to fully utilize hardware resources such as GPUs and SIMD units. As a use case, we run mpEDM on AI Bridging Cloud Infrastructure (ABCI) using datasets of an entire animal brain sampled at single neuron resolution to identify dynamical causation patterns across the brain. mpEDM is 1, 530× faster than cppEDM and a dataset containing 101, 729 neuron was analyzed in 199 seconds on 512 nodes. This is the largest EDM causal inference achieved to date.

DOI： 10.1109/ICPADS51040.2020.00035

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/icpads/icpads2020.html#Watanakeesuntorn20

Language：Japanese   Publishing type：Research paper (scientific journal)   Publisher：Japan Society for Software Science and Technology  

AiBiC Kansai offers an education course, AiBiC Spiral, designed to develop abilities to solve individual social issues using Big Data, artificial intelligence, and Cloud technologies through teamwork. In this paper, we report on the problems of the environments for the education program and the discrepancies in the expectations on topics and contents found by the analysis of the questionnaire conducted in 2017, and describe the activities and results performed to improve the program in 2018. Through the improvements, we have found that the problem on the program environments has been dissolved. However, as for the discrepancies in the expectations, we have revealed that only some part of the problem was improved and there were still some students found the topics and contents provided by the program were different from their expectations.

DOI： 10.11309/jssst.37.1_19

Scopus

CiNii Books

J-GLOBAL

researchmap

Language：English   Publishing type：Research paper (scientific journal)   Publisher：IEICE-INST ELECTRONICS INFORMATION COMMUNICATIONS ENG  

The success and scale of the Internet and its protocol IP has spurred emergent distributed technologies such as fog/edge computing and new application models based on distributed containerized microservices. The Internet of Things and Connected Communities are poised to build on these technologies and models and to benefit from the ability to communicate in a peer-to-peer (P2P) fashion. Ubiquitous sensing, actuating and computing implies a scale that breaks the centralized cloud computing model. Challenges stemming from limited IPv4 public addresses, the need for transport layer authentication, confidentiality and integrity become a burden on developing new middleware and applications designed for the network's edge. One approach - not reliant on the slow adoption of IPv6 - is the use of virtualized overlay networks, which abstract the complexities of the underlying heterogeneous networks that span the components of distributed fog applications and middleware. This paper describes the evolution of the design and implementation of IP-over-P2P (IPOP) - from its purist P2P inception, to a pragmatic hybrid model which is influenced by and incorporates standards. The hybrid client-server/P2P approach allows IPOP to leverage existing robust and mature cloud infrastructure, while still providing the characteristics needed at the edge. IPOP is networking cyber infrastructure that presents an overlay virtual private network which self-organizes with dynamic membership of peer nodes into a scalable structure. IPOP is resilient to partitioning, supports redundant paths within its fabric, and provides software defined programming of switching rules to utilize these properties of its topology.

DOI： 10.1587/transcom.2019CPI0001

Web of Science

Scopus

researchmap

Language：English   Publishing type：Research paper (other academic)  

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

Data centers are centralized facilities where computing and networking hardware are aggregated to handle large amounts of data and computation. In a data center, computing resources such as CPU and memory are usually managed by a resource manager. The resource manager accepts resource requests from users and allocates resources to their applications. A commonly known problem in resource management is that users often request more resources than their applications actually use. This leads to the degradation of overall resource utilization in a data center. This paper aims to improve resource utilization in data centers by predicting the required resource for each application. We designed and implemented a neural network model based on Long Short-Term Memory (LSTM) to predict more efficient resource allocation for a job based on historical data. Our model has two LSTM layers each of which learns the relationship between: (1) allocation and usage, and (2) CPU and memory. We used Googles cluster-usage trace, which contains a trace of resource allocation and usage for each job executed on a Google data center, to train our neural network. Googles cluster scheduler simulator was used to evaluate our proposed method. Our simulation indicated that the proposed method improved the CPU utilization and memory utilization by 10.71% and 47.36%, respectively, compared to a conventional resource manager. Moreover, we discovered that increasing the memory cell size of our LSTM model improves the accuracy of the prediction in return for longer training time.

DOI： 10.1109/CLUSTER.2019.8891022

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/cluster/cluster2019.html#ThonglekITIN19

Language：Japanese  

CiNii Books

researchmap

Language：Japanese   Publishing type：Research paper (scientific journal)  

CiNii Books

J-GLOBAL

researchmap

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper (scientific journal)  

CiNii Books

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper (conference, symposium, etc.)  

J-GLOBAL

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

An innovative engineer who can address a social challenge using big data processing, AI and cloud computing technologies with the generation of new business and value is required from industry. enPiT is an education project to develop the advanced IT engineer based on practical education in cooperation between industry and academia promoted by Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan. In this paper, we introduce how we designed PBL centered curriculum named AiBiC Spiral under the framework of enPiT education project, and analyze educational effect of our program based on work products and the questionnaire result targeting students who took and completed to the program of 2017.

DOI： 10.1109/BCD2018.2018.00013

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/bcd/bcd2018.html#SaikiFIKNMYK18

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

In this paper, we propose an ASP system for shopping path analysis and a cloud based system with the necessary analysis service for marketing strategies that combines sales data with shopping path data. This paper explains a recommendation system based on position information or an application that calculates a basic marketing indicator using shopping path data by introducing a framework of ASP for shopping path analysis. Because the proposed ASP service is built on a cloud based system, the users can easily access the service through the web based interface and perform large-scale data processing using the computational resources of the cloud based system at low cost.

DOI： 10.1109/APWConCSE.2017.00017

Web of Science

Scopus

researchmap

Publishing type：Research paper (international conference proceedings)  

Distributed denial of service (DDoS) attack is one of the most costly attacks for IT system in terms of time and money. In this paper, the use of machine learning algorithms for DDoS detection has been addressed. The traditional SVM and new emerging deep learning algorithm, namely Deep Feed Forward (DFF), are evaluated. The DARPA Scalable Network Monitoring and DARPA 2009 DDoS attacks dataset is used to test the effectiveness of these two algorithms. The dataset is preprocessed to find the potential speedup of the classification process. From the experiments, DFF deep learning algorithm has achieved a high accuracy of 99.63% with the training time of 289.614 secs. For SVM, the highest accuracy achieved is 93.01%, with the training time of 371.118 secs. Anyway, SVM is able to deliver a faster classification time. Therefore, DFF is suitable for the situation when accuracy is the main concern while SVM can be used when speed of classification is a critical factor.

DOI： 10.1109/ICSEC.2018.8712757

Scopus

researchmap

Language：English   Publishing type：Research paper (other academic)  

researchmap

Language：English  

CiNii Books

researchmap

Other Link： https://ndlsearch.ndl.go.jp/books/R000000004-I028943716

Language：English  

CiNii Books

researchmap

Other Link： https://ndlsearch.ndl.go.jp/books/R000000004-I028942930

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE Computer Society  

Software Defined Network (SDN) is a new approach of networking that emphasis on the network programmability and more dynamic control of the network. OpenFlow is the broadly used standard to implement the SDN network. However, understanding the dynamic behavior of an OpenFlow network is still challenging since the information about the operation is distributed across numerous network switches. This paper presents a tool call Opimon (OpenFlow Interactive Monitoring) that monitor and visualizes the OpenFlow network based SDN. Opimon provides a real-Time monitoring and visualization of an OpenFlow network including network topology and flow tables of each switch. Web based user interface enables a user to quickly follows the network behavior and identify the problem. Opimon has been deployed successfully on a PRAGMA-ENT international testbed and enables researchers to understand more about the behavior of the testbed.

DOI： 10.1109/ICTKE.2017.8259622

Scopus

researchmap

Language：Japanese   Publisher：電子情報通信学会  

CiNii Books

J-GLOBAL

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

Since the Internet is an aggregation of multiple ASes (Autonomous Systems), congestion control and utilization are not globally optimized. For example, it is not uncommon that a direct shortest route with low latency delivers less bandwidth than an alternative, long and roundabout route. Previous research has shown that geospatially distributed computing instances in commercial clouds offer users an opportunity to deploy relay points to detour potentially congested ASes, and as a means to diversify paths to increase overall bandwidth and reliability. Such opportunity comes with a cost, as cloud-routed paths incur cost of not only provisioning of computing resources, but also for additional traffic to/from Internet. Well-established protocols, such as TCP, were created based on assumption of single end-point to end-point transfer; nonetheless, current computing devices have multiple end-points, and the increasing availability of overlay networks allows multiplexing multiple virtual network flows into a single physical network interface. In this paper, we empirically evaluate the extent to which using cloud paths to transfer data in parallel with the default Internet path can improve the end-to-end bandwidth in bulk data transfers. In our evaluation, we consider single-stream and multi-stream TCP transfers across one or more paths. Moreover, we suggest an application level design pattern that takes advantage of this improved aggregate bandwidth to reduce data transfer times.

DOI： 10.1109/CloudCom.2017.27

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/cloudcom/cloudcom2017.html#JeongFI17

Language：Japanese   Publisher：日本ソフトウェア科学会  

CiNii Books

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

CiNii Books

researchmap

Other Link： http://id.nii.ac.jp/1001/00183236/

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

Multi-tenant data centers for cloud computing require the deployment of virtual private networks for tenants in an on-demand manner, providing isolation and security between tenants. To address these requirements, network virtualization techniques such as encapsulation and tunneling have been widely used. However, these approaches inherently incur processing overhead on end-points (such as the host hypervisor), reducing the effective throughput for the tenant virtual network compared to the native network. This problem is exacerbated with increases in line rates, now exceeding 10Gbps. In this paper, we introduce PARES (PAcket REwriting on SDN), a novel technique which uses the packet rewriting feature of SDN switches to provide multi-tenancy in data center networks at edge switches, thereby reducing the load on end-point hypervisors and improving the throughput, compared to tunneling. Experiments in an SDN testbed show that our proposed data center arhictecture with PARES achieves near line-rate multi-tenancy virtualization with 10Gbps links (compared to 20% of line-rate for VXLAN tunneling), without incurring processing overhead at end-point hypervisors or guest servers. Additionally, the paper evaluates the scalability of PARES for ARP protocol handling and with respect to number of SDN flow entries.

DOI： 10.1109/CLOUD.2017.11

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/IEEEcloud/IEEEcloud2017.html#JeongFI17

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

Similar to Virtualization, Linux Containers (LXC) provides high-performance, lightweight computing resource allocation and isolation. Each LXC container has a resource overhead smaller than that of a virtual machine, leading to significantly lower container migration time and making frequent container placement modification a viable optimization technique. Traditional container scheduling mechanisms do not leverage this property of LXC. Generally, a scheduler tries to find the most optimal placement for a new container, the allocated host then executes the scheduled container until the end of the container's life cycle. This strategy works fine for short-lived containers. With a long-lived container such as a server process becoming more and more common, and the container placement calculated at the beginning of the execution may not remain optimal during the container's lifetime, since the other containers are moving in and out of the cluster. This research proposes container rebalancing, a novel scheduling mechanism with a rebalancing process working alongside a scheduling process. The container rebalancing method increases LXC cluster utilization while maintaining minimal interference with the scheduling process. This is done by continuously modifying container placement, by using the rebalancing process, in order to load-balance utilization of each host in the LXC cluster. LXC cluster simulation driven by Google's cluster data is used to verify the feasibility of container rebalancing. Simulation results show an observable increase in container scheduled rate and cluster utilization with no drawback, suggesting that container rebalancing is a promising method.

DOI： 10.1109/COMPSAC.2017.94

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/compsac/compsac2017-1.html#U-ChupalaWIDI17

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：Institute of Electrical and Electronics Engineers Inc.  

Compared to traditional head-mounted displays (HMD) which are mostly for industrial usage, the more compact and affordable HMDs for mass-market have claimed to be an emerging impact in the gaming industry. Such kind of HMDs, which are represented by Oculus Rift, HTC Vive and Playstation VR, have proved to be a more accessible solution for game players to experience VR gaming. However, the popularity of VR gaming is still hindered by the requirement of relatively heavy local setup in order to maintain smooth VR experience. With this in consideration, in this paper we provide our vision on improving the accessibility of VR experience by utilizing Cloud Gaming technique. Furthermore, we will also discuss the expected difficulties to be encountered while applying such technique.

DOI： 10.1109/ISUVR.2017.15

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/isuvr/isuvr2017.html#ChanIWI17

Language：English   Publishing type：Research paper (scientific journal)   Publisher：Springer Science and Business Media LLC  

DOI： 10.1007/s40012-017-0171-y

researchmap

Other Link： http://link.springer.com/content/pdf/10.1007/s40012-017-0171-y.pdf

Language：English   Publishing type：Research paper (scientific journal)   Publisher：SERIOUS GAMES SOC  

The emerging Cloud Gaming Service provides a highly accessible video gaming experience. With Cloud Gaming, potential players without enough local resource can access high-quality gaming using low-spec devices. With advancing technology, we consider that if the processing power at low-spec devices can be well harvested, the quality delivered on Cloud Gaming can be further improved. Therefore, we propose a Hybrid-Streaming System that aimed at improving the graphic quality delivered by Cloud Gaming. By utilizing the available rendering power from both the Cloud Server and client PC, the system distributes rendering operations to both sides to achieve the desired improvement. Quantitative results show the proposed method improves graphics quality, as well as reducing the server's workload while attaining acceptable network bandwidth consumption levels.

DOI： 10.17083/ijsg.v4i2.163

Web of Science

researchmap

Other Link： https://dblp.uni-trier.de/db/journals/ijsg/ijsg4.html#ChanIWUI17

Language：English   Publishing type：Research paper (other academic)  

researchmap

Language：English   Publishing type：Research paper (scientific journal)   Publisher：Information Processing Society of Japan  

A large amount of data needs to be transferred from one site to another as fast as possible in the computational science fields. To achieve high-speed data transfer, many applications utilize multiple TCP streams. However, since multiple TCP streams of applications are usually routed according to the default IP routing protocol, only a single shortest path among the multiple paths can be utilized for the data transfer. This research proposes a multipath controller that increases the performance of data transfer by leveraging multiple paths simultaneously for parallel TCP streams. For this purpose, we utilize the Software-Defined Networking (SDN) technology and its implementation, OpenFlow. Furthermore, we proposed a method to determine optimal numbers of parallel TCP streams to be assigned for each path according to its own network condition. This paper presents the design and implementation of the proposed system. As a case study, we applied our proposed system on GridFTP and evaluated the performance improvement. The results demonstrate that our proposed system accelerates the data transfer of GridFTP in both a virtual and a real global-scale environment.

DOI： 10.2197/ipsjjip.25.924

Scopus

researchmap

Language：English   Publishing type：Research paper (scientific journal)   Publisher：WILEY  

Multipath transmission control protocol(TCP), or MPTCP, is a widely-researched mechanism that allows a single application-level connection to be split to more than 1 TCP stream and, consequently, more than 1 network interface, as opposed to the traditional TCP/IP model. Being a transport layer protocol, MPTCP can easily interact between the application using it and the network supporting it. However, MPTCP does not have control of its own route. Default IP routing behavior generally takes all traffic through the shortest or best metric path. However, this behavior may actually cause paths to collide with each other, creating contention for bandwidth in a number of edges. This can result in a bottleneck that limits the throughput of the network. Therefore, a multipath routing mechanism is necessary to ensure smooth operation of MPTCP. We created smoc, a simple multipath OpenFlow controller, that uses only topology information of the network to avoid collision where possible. Evaluation of smoc in a virtual local-area and a physical wide–area software-defined networks showed favorable results as smoc provided better performance than simple or spanning tree–routing mechanisms.

DOI： 10.6084/M9.FIGSHARE.1558361.V1

Web of Science

Scopus

arXiv

researchmap

Language：English   Publishing type：Research paper (scientific journal)   Publisher：WILEY  

The Pacific Rim Application and Grid Middleware Assembly (PRAGMA) is an international community of researchers that actively collaborate to address problems and challenges of common interest in eScience. The PRAGMA Experimental Network Testbed (PRAGMA-ENT) was established with the goal of constructing an international software-defined network (SDN) testbed to offer the necessary networking support to the PRAGMA cyberinfrastructure. PRAGMA-ENT is isolated, and PRAGMA researchers have complete freedom to access network resources to develop, experiment, and evaluate new ideas without the concerns of interfering with production networks. In the first phase, PRAGMA-ENT focused on establishing an international L2 backbone. With support from the Florida Lambda Rail, Internet2, PacificWave, Japan Gigabit Network, and TaiWan Advanced Research and Education Network, PRAGMA-ENT backbone connects openflow-enabled switches at University of Florida, University of California, San Diego, Nara Institute of Science and Technology (Japan), Osaka University (Japan), National Institute of Advanced Industrial Science and Technology (Japan), and National Applied Research Laboratories (Taiwan). The second phase of PRAGMA-ENT consisted of an evaluation of technologies for the control plane that enables multiple experiments (ie, OpenFlow controllers) to coexist. Preliminary experiments with FlowVisor revealed some limitations leading to the development of a new approach, called AutoVFlow. This paper describes our experience in the establishment of PRAGMA-ENT backbone (with international L2 links), its current status, and plans for the control plane. Discussion of preliminary application ideas, including optimization of routing control; multipath routing control; extending the backbone using overlay network; and remote visualization are also discussed.

DOI： 10.1002/cpe.4138

Web of Science

Scopus

researchmap

Language：Japanese  

J-GLOBAL

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

Software Defined Network (SDN) is a new approach of networking that emphasis on the network programmability and more dynamic control of the network. OpenFlow is the broadly used standard to implement the SDN network. However, understanding the dynamic behavior of an OpenFlow network is still challenging since the information about the operation is distributed across numerous network switches. This paper presents a tool call Opimon (OpenFlow Interactive Monitoring) that monitor and visualizes the OpenFlow network based SDN. Opimon provides a real-time monitoring and visualization of an OpenFlow network including network topology and flow tables of each switch. Web based user interface enables a user to quickly follows the network behavior and identify the problem. Opimon has been deployed successfully on a PRAGMA-ENT international testbed and enables researchers to understand more about the behavior of the testbed.

Web of Science

researchmap

Language：Japanese  

researchmap

Language：Japanese  

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

We propose an extension of distcloud-fs for achieving a global distributed POSIX file system in intercloud environment. The file system works in intercloud environment, that is, on the top of multiple independent cloud services. The important capability of the intercloud extension is that there is no need for users to take care of infrastructure layer, while the file system stores data into multiple cloud services, transparently and automatically. As the result, cloud users may have diversity about vendors and locations for data platform, as discussed in various intercloud research projects. In this paper, we describe the architecture of the distcloud-fs and its intercloud extension. We also describe design and implementation of the prototype and discuss some typical use cases of the file system with practical demonstrations.

DOI： 10.1109/IIAI-AAI.2016.75

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/iiaiaai/iiaiaai2016.html#NakagawaKSIKKKY16

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

We demonstrate high-speed data transfer GridFTP using a multipath control mechanism with SDN (Software-Defined Networking). GridFTP is a typical tool that has been developed and widely used for bulk data transfer over a wide area network in the field. GridFTP supports a parallel high-speed data transfer scheme using multiple TCP streams. However, one of the shortest paths is used solely for data transfer in the default IP routing while there are multiple network paths (multipath) exist between widely-distributed sites. In this study, we propose a system that distributes the parallel TCP streams of GridFTP into multiple network paths by a traffic engineering technique brought by SDN. Our proposed system has achieved approximately 20% better performance than the conventional method in the best case in a global-scale real enviroment.

DOI： 10.1109/ICDCS.2016.80

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/icdcs/icdcs2016.html#HuangNII16

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

Nowadays, users' computation requests to a high-performance computing (HPC) environment have been increasing and diversifying for requiring large-scale simulations and analysis in the various science fields. In order to efficiently and flexibly handle such computation requests, resource allocation of the virtualized computational resources on an HPC cluster system such as Cloud Computing service is attracting attention. Currently, we aim to realize a novel resource management system (RMS) that enable to handle various resources of an HPC cluster system, and have been studying and developing the SDN-enhanced Job Management System (JMS) Framework, which can manage an interconnect as network resources by integrating Software Defined Networking (SDN) concept into a traditional JMS. However, the current SDN-enhanced JMS Framework cannot allocate virtualized computational resources to a job because the computational resource management is performed by the mechanism of a traditional JMS. In this paper, we propose a mechanism to handle virtualized computational resources on the SDN-enhanced JMS Framework. This mechanism enables to deploy virtual machines (VMs) requested by the user to the computing nodes allocated to a job and execute job's processes in the VMs.

DOI： 10.1109/MIPRO.2016.7522145

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/mipro/mipro2016.html#WatashibaDAIKYK16

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

The MPI reduction operation such as MPI-Reduce and MPI-Allreduce are frequently used and time-consuming operations. The performance enhancement of these operations can substantially speed up large-scale parallel applications. In this paper, a greedy based MPI-Reduce algorithm called Greedy Shortest Binomial Tree (GSBT) is proposed. This proposed algorithm leverages SDN technology and OpenFlow network to speed up MPI reduction operations. This is accomplished using network topology information from the OpenFlow controller to reduce overall hops in message transmission. The implementation of the proposed algorithm by modifying MPI library and OpenFlow controller is presented. The proposed GSBT algorithm has been evaluated in a real test-bed to compare with the traditional approaches used in both MPICH and Open MPI. The result shows that GSBT algorithm is faster than standard algorithms 30.48-66.35% for Open MPI and faster 50.77-82.89% for MPICH when message size between 2 KB - 24 KB.

DOI： 10.1109/ISCIT.2015.7458357

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/iscit/iscit2015.html#MakpaisitIUDTK15

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

High performance computing is required for Big Science application because the proliferation and huge amount of scientific data that needs to be analyzed is a serious problem. Traditionally, network resources were generally assumed as a static resource users cannot control on demand. By integrating network programmability to every stage of a scientific workflow, this study explores a next-generation high performance computing infrastructure where both computational and network resources are flexibly sliced and efficiently leveraged based on the resource requirements of the scientific applications. Technically, Software Defined Networking has been adopted as a key technology for this purpose. In this paper the concept and goals of a next-generation high performance computing infrastructure is introduced and the current status of our research is discussed.

DOI： 10.1109/ICCCRI.2015.13

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/icccri/icccri2015.html#DateAKTKWUIYKS15

Language：English   Publishing type：Research paper (other academic)  

researchmap

Language：English   Publishing type：Research paper (scientific journal)   Publisher：Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)  

The emerging Cloud Gaming Service provides highly accessible video gaming experience. However, in term of the gaming quality, Cloud Gaming is not competitive to rival with traditional gaming because of network constraints. Especially, 3D game contents streamed as encoded video sequence is suitable in a network environment, but the resulted lower graphic quality may not meet client’s demand. Therefore, we propose a Hybrid-Streaming System that aims at improving graphic quality delivered on Cloud Gaming. By utilizing available rendering power from both Cloud Server and client’s PC, the system distributes rendering operations to both sides to achieve the desired improvement. Quantitative results show the improvement of graphic quality from the proposed method, as well as reducing server’s workload and attaining acceptable network bandwidth consumption.

DOI： 10.1007/978-3-319-46100-7_13

Web of Science

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：ELSEVIER  

Virtual high-throughput biochemical screening offers a cost-effective alternative to the empirical testing of millions of compounds. However, virtual screening data often requires some manual processing of the data to eliminate false positives, evaluate the ligand-macromolecule fit, and identify new molecular interactions. This analysis is generally hindered by highly specific software and hardware requirements and complex user interfaces. Hydra is an HTML5 and JavaScript based application, which ameliorates this issue by displaying ligand-macromolecule models calculated by virtual screening programs in a single, simple online interface. The application is capable of loading raw data sets from the DOCK virtual screening platform and utilizing pre-processed datasets from other software to display compounds side-by-side in a user-defined size grid of 3Dmol.js instances. It also searches databases for selected compound information to natively display within the interface. This tool provides a highly accessible platform for streamlined virtual screening results analysis.

DOI： 10.1016/j.procs.2016.05.427

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/iccS/iccS2016.html#SeraMWIH16

Language：English   Publishing type：Research paper (scientific journal)   Publisher：ELSEVIER  

To empower scientists who are engaged in nation-wide or global-scale collaborative projects for scientific discovery, a large amount of scientific data needs to be visualized and then shared among the scientists. Tiled Display Wall (TDW) has been widely accepted and used for visualization of large-scale scientific data. Scalable Adaptive Graphics Environment (SAGE) has received attention from scientists as a middleware that organizes multiple display monitors into a network-aware large display monitor. Using a SAGE TDW, scientists can display multiple visualized contents on a single display monitor, each of which can be located at geographically distant site managed by other organizations. However, SAGE does not have a mechanism for managing multiple visualized data streams heading to a single TDW. In a conventional network, data flows for a same destination tend to share a same link, resulting in drop of packets and therefore poor visual quality. Moreover, because of the flexible nature of SAGE, rate of each visual data flow may change dynamically as a result of user interaction on a TDW, such as moving and resizing an application window. For the reason above, we propose and develop a dynamic route allocation method that switches packet flows onto network links where better performance is expected, in response to user interaction such as window movement and resizing. Technically, we have leveraged OpenFlow, an implementation of Software Defined Networking (SDN), to integrate network programmability into SAGE. In this paper, we show how SAGE enhanced with the proposed method succeeded in avoiding network congestion and improving the quality of visualization on the TDW over the wide area OpenFlow network on the Internet.

DOI： 10.1016/j.future.2015.09.033

Web of Science

Scopus

CiNii Books

researchmap

Language：Japanese   Publishing type：Research paper (scientific journal)  

It is important to ensure robustness and availability of information infrastructures against disasters and faults. There are many technologies of distribution and redundancy, but most of them do not provide practical validation methods on faults. We propose a platform to evaluate disaster tolerance and fault resistance of distributed systems with an SDN technology emulating artificial disasters and faults. In this paper, we introduce a design and its implementation of the platform, and then also show the effectiveness of the platform through evaluation experiments on it.

CiNii Books

J-GLOBAL

researchmap

Language：Japanese   Publisher：電子情報通信学会  

CiNii Books

J-GLOBAL

researchmap

Language：Japanese   Publisher：電子情報通信学会  

CiNii Books

J-GLOBAL

researchmap

Language：English   Publishing type：Research paper (scientific journal)   Publisher：BENTHAM SCIENCE PUBL LTD  

The drug discovery process in general is a very resource intensive undertaking that has existed for a very long time. In the last two decades, performing molecular simulations that determine the level of interaction between a protein and ligand have been refined to the point where they are now an essential part of the drug discovery process. These simulations serve to reduce the time to discovery and improve the positive “hit” rates when screening for molecule with biological activity. As a result, the chemical search space is greatly reduced in silico, prior to any in vitro experiments that validate the results. Recently, there have been many advances in computer science technologies that have improved the virtual screening process. This paper will give a brief overview of the virtual screening process and then summarize the current state-of-the-art technologies applied to virtual screenings. Both biomedical researchers and computer scientists can use this review as a guide to the implementation requirements for computational resources of virtual screening.

DOI： 10.2174/1381612822666160414142530

Web of Science

Scopus

PubMed

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

With the constant increase in the number and variety of small molecule chemical compounds, drug discovery is becoming a very resource intensive endeavor. Performing molecular simulations of ligand-protein binding by virtual screening has become an integral part of the discovery process. Cloud computing is an efficient choice to execute these large-scale screenings, given that large compute allocations are not accessible to many researchers. This research focused on developing a multi-site cloud environment that combines small allocations of virtual machines in multiple locations connected through a virtual networking system (ViNe), and compared two parallelization approaches: Message Passing Interface (MPI) and MapReduce using Hadoop. Virtual screenings were conducted using DOCK, a protein-ligand molecular interaction simulation program. Multiple DOCK test simulations through MPI and Hadoop were run to assess the performance and flexibility of the environment. These tests indicated that MPI and MapReduce offer comparable scalability performance, and that network latency has a significant influence on low accuracy simulations. Furthermore, differences in performance at individual cloud resource sites were reduced on average because of the larger combined pool of resources. This project prototyped and assessed a fully functional multi-site cloud environment for virtual screenings, which can be used to guide small laboratories in deploying their own cloud-based screenings.

DOI： 10.1109/eScience.2015.49

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/eScience/eScience2015.html#NguyenMTDIH15

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

A large amount of scientific data needs to be transferred from one site to another as fast as possible in the computational science fields. High-speed data transfer between sites is very important, especially in the Grid computing field, GridFTP has been widely used for bulk data transfer over a wide area network. GridFTP achieves greater performance by supporting parallel TCP streams. Using parallel TCP streams improves the throughput of slow-start algorithms and lossy networks even on a single path. This research proposes a traffic engineering technique that increases the data transfer performance by using multiple paths simultaneously for the parallel TCP streams. For this purpose, we use Software-Defined Network (SDN) technology and its implementation, OpenFlow. This paper presents the design and implementation of the proposed system. Our performance evaluation demonstrates that our proposed system can accelerate GridFTP Transfer in both virtual and real global-scale environments.

DOI： 10.1109/eScience.2015.37

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/eScience/eScience2015.html#HuangNII15

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

Bandwidth and latency are two major factors that contribute the most to network application performance. Between each pair of switches in a network, there may be multiple paths connecting them. Each path has different properties because of multiple factors. Traditional shortest-path routing does not take this knowledge into consideration and may result in sub-optimal performance of applications and underutilization of network. We proposed a concept of 'bandwidth and latency aware routing'. The idea is that we could improve overall performance of the network by separating application into bandwidth-oriented and latency-oriented application and allocate different route for each type of application accordingly. We also proposed a design of this network system implemented using Open Flow. Routes are calculated from monitored information using Dijkstra algorithm and its variation. To support our design, we show a use case in which our design performs better than traditional routing as well as evaluation results.

DOI： 10.1109/CloudCom.2014.90

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/cloudcom/cloudcom2014.html#U-ChupalaIIKUDAYK14

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

In the era of cloud computing, data centers that accommodate a series of user-requested jobs with a diversity of resource usage pattern need to have the capability of efficiently distributing resources to each user job, based on individual resource usage patterns. In particular, for high-performance computing as a cloud service which allows many users to benefit from a large-scale computing system, a new framework for resource management that treats not only the CPU resources, but also the network resources in the data center is essential. In this paper, an SDN-enhanced JMS that efficiently handles both network and CPU resources and as a result accelerates the execution time of user jobs is introduced as a building block technology for such a HPC cloud. Our evaluation shows that the SDN-enhanced JMS efficiently leverages the fat-tree interconnect of cluster systems running behind the cloud to suppress the collision of communications generated by different jobs.

DOI： 10.1109/CloudCom.2014.82

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/cloudcom/cloudcom2014.html#WatashibaDAKIYKST14

Publishing type：Research paper (scientific journal)  

The Pacific Rim Application and Grid Middleware Assembly (PRAGMA) is an
international community of researchers that actively collaborate to address
problems and challenges of common interest in eScience. The PRAGMA Experimental
Network Testbed (PRAGMA-ENT) was established with the goal of constructing an
international software-defined network (SDN) testbed to offer the necessary
networking support to the PRAGMA cyberinfrastructure. PRAGMA-ENT is isolated,
and PRAGMA researchers have complete freedom to access network resources to
develop, experiment, and evaluate new ideas without the concerns of interfering
with production networks.
In the first phase, PRAGMA-ENT focused on establishing an international L2
backbone. With support from the Florida Lambda Rail (FLR), Internet2,
PacificWave, JGN-X, and TWAREN, PRAGMA-ENT backbone connects Open\-Flow-enabled
switches at University of Florida (UF), University of California San Diego
(UCSD), Nara Institute of Science and Technology (NAIST, Japan), Osaka
University (Japan), National Institute of Advanced Industrial Science and
Technology (AIST, Japan), and National Center for High-Performance Computing
(Taiwan).
The second phase of PRAGMA-ENT consisted of evaluation of technologies for
the control plane that enables multiple experiments (i.e., OpenFlow
controllers) to co-exist. Preliminary experiments with FlowVisor revealed some
limitations leading to the development of a new approach, called AutoVFlow.
This paper will share our experience in the establishment of PRAGMA-ENT
backbone (with international L2 links), its current status, and control plane
plans. Discussion on preliminary application ideas, including optimization of
routing control; multipath routing control; and remote visualization will also
be discussed.

arXiv

researchmap

Other Link： https://dblp.uni-trier.de/db/journals/corr/corr1509.html#IchikawaTHYLKUH15

Publishing type：Research paper (scientific journal)  

Multipath transmission control protocol(TCP), or MPTCP, is a widely-researched mechanism that allows a single application-level connection to be split to more than 1 TCP stream and, consequently, more than 1 network interface, as opposed to the traditional TCP/IP model. Being a transport layer protocol, MPTCP can easily interact between the application using it and the network supporting it. However, MPTCP does not have control of its own route. Default IP routing behavior generally takes all traffic through the shortest or best metric path. However, this behavior may actually cause paths to collide with each other, creating contention for bandwidth in a number of edges. This can result in a bottleneck that limits the throughput of the network. Therefore, a multipath routing mechanism is necessary to ensure smooth operation of MPTCP. We created smoc, a simple multipath OpenFlow controller, that uses only topology information of the network to avoid collision where possible. Evaluation of smoc in a virtual local-area and a physical wide–area software-defined networks showed favorable results as smoc provided better performance than simple or spanning tree–routing mechanisms.

DOI： 10.1002/cpe.4134

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/journals/corr/corr1509.html#NakasanIIU15

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

In the wide-area distributed systems such as the Internet, it is important to ensure the robustness and the availability through data distribution and redundancy for disaster and failure. We propose a platform which called SDDE (Software Defined Disaster Emulation) to evaluate disaster resistance and disaster tolerance by emulating some disaster and failure. In this paper, we report an evaluation experiment of the disaster resistance for wide-area distributed storage system using our proposed SDDE platform.

CiNii Books

J-GLOBAL

researchmap

Language：Japanese  

J-GLOBAL

researchmap

Language：Japanese  

J-GLOBAL

researchmap

Language：Japanese  

CiNii Books

J-GLOBAL

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

In the wide-area distributed systems such as the Internet, it is important to ensure the robustness and the availability through data distribution and redundancy for disaster and failure. We propose a platform which called SDDE (Software Defined Disaster Emulation) to evaluate disaster resistance and disaster tolerance by emulating some disaster and failure. In this paper, we report an evaluation experiment of the disaster resistance for BGP routing system using our proposed SDDE platform.

CiNii Books

J-GLOBAL

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

ICT systems are expected to ensure high availability against not only possible failures but also unexpected multiple failures due to wide-scale disaster. It is necessary to continuous evaluations on the operation through education or training. This study proposes a verification platform which enables to arbitrary-generate disasters or failures using SDN to evaluate disaster tolerance of the distributed system.

CiNii Books

J-GLOBAL

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

It is important to ensure toughness and robustness of infrastructure for information network such as the Internet. To ensure toughness and robustness, there are several solution such as redundant configurations, SPoF (single point of failure) less wide-are distribution. In this paper, authors design and evaluate a platform to evaluate disaster tolerance of communication systems by immitating existing disaster records.

CiNii Books

J-GLOBAL

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

It is important to ensure toughness and robustness of infrastructure for information network such as the Internet. In this paper, authors validate and evaluate distributed systems by using a SDDE (Software Defined Disaster Emulation) platform.

CiNii Books

J-GLOBAL

researchmap

Language：English  

CiNii Books

researchmap

Publishing type：Research paper (scientific journal)   Publisher：Springer Science and Business Media LLC  

DOI： 10.1007/s12626-014-0045-9

researchmap

Other Link： http://link.springer.com/article/10.1007/s12626-014-0045-9/fulltext.html

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

In recent years, much attention has been paid to wide area distributed storages to backup data remotely and ensure that business processes can continue in terms of disaster recovery. In the 'distcloud' project, authors have been involved in the research of wide area distributed storage by clustering many computer resources located in geographically distributed areas, where the number of sites is more than 2 (N>2). The storage supports a shared single POSIX file system so that LDLM (Long Distance Live Migration) of VMs (Virtual Machines) works well between multiple sites. We introduce the concept and basic architecture of the wide area distributed storage and its technical improvement for LDLM. We describe the result of our experiment, that is, 1) Nation Wide Live Migration (about 500Km) in Japan, and 2) Transpacific Live Migration (over 24,000Km). We show the technical benefit of the current implementation and discuss suitable applications and remaining issues for further research topics.

DOI： 10.1109/COMPSAC.2014.71

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/compsac/compsac2014.html#NakagawaIKKKS14

Language：English   Publisher：Information Processing Society of Japan (IPSJ)  

This paper discusses the use of Multipath TCP (MPTCP), which is a TCP extension that allows multiple TCP flows to be associated to one application-layer logical connection, coupled with OpenFlow traffic engineering in a single stack to provide a comprehensive multipathing solution, with OpenFlow providing optimal path sets while MPTCP utilizing them. This design should be able to maximize bandwidth and network path utilization by allowing hosts to take advantage of presently-unused paths. Design of the testbed, also to be used by our research group in future projects, is also discussed in this paper. Finally, we discuss evaluation of network performance when using multiple paths, as well as concerns raised by our work. In summary, our system functioned as expected and provided feasible performance in small virtual network. This design should be scalable to benefit distributed file storage systems, data-intensive services, or any high-performance computing systems.

CiNii Books

researchmap

Language：English   Publisher：Information Processing Society of Japan (IPSJ)  

Bandwidth and latency are two major factors that contribute the most to network application performance. Between each pair of switches in a network, there may be multiple paths connecting them. Each path has different properties because of multiple factors. Traditional shortest-path routing does not take this knowledge into consideration and may result in sub-optimal performance of applications and underutilization of network. We proposed a concept of "bandwidth and latency aware routing". The idea is that we could improve overall performance of the network by separating application into bandwidth-oriented and latency-oriented application and allocate different route for each type of application accordingly. We also proposed a design of this network system implemented using OpenFlow. Routes are calculated from monitored information using Dijkstra algorithm and its variation. To support our design, we show a use case in which our design performs better than traditional routing.

CiNii Books

researchmap

Language：Japanese   Publishing type：Research paper (conference, symposium, etc.)  

The authors have been researching and developing widely distributed storage environment that is characterized by high performance of random I/O access. It is difficult for conventional distributed storage to serve practicable performance for using virtualized machines, especially on live migration. In this paper, the authors show the latest result of performances of I/O on proposed distributed algorithm. A result of 24 thousands km live-migration demonstrated in SC13 (Supercomputer Conference 2013) is also shown.<br />

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：Institute of Electrical and Electronics Engineers Inc.  

In this work, a network benchmarking system that measures the link performance in a large network system is presented. Our previous work on monitoring as a Service system (MaaS) is extended to create a monitoring infrastructure that is capable of regular benchmarking and maintaining network performance graph. The benchmark traffic conflict is avoided by partitioning the large measured network into many smaller subnetworks that can be measured separately. A smart scheduler is developed to control the measurement and ensure that a more accurate bandwidth and latency measurement is performed. The system is developed to gather information that used to manage large Software Defined Network in the future.

DOI： 10.1109/ICSEC.2014.6978126

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：SPRINGER-VERLAG BERLIN  

Collective communication is of great importance in MPI because the execution time of an MPI program is affected by the communication performance it can gain. Particularly these days, when a cluster system composed of multiple computing nodes has become dominant as a large-scale computing system, the execution time of collective communication affects the total execution time of the MPI program. However, in many implementations of MPI, collective communication is developed to make use of unicast-based communication in a repeated and combined way, which may result in inefficient communication. In this paper, we explore the use of a Software-Defined Network, which was originally expected to help network administrators operate networks through central control in a software-programming manner, to accelerate MPI-Bcast, a basic collective communication used in MPI. The evaluation in this paper indicates that our prototyped SDN-MPI-Bcast is superior to MPI-Bcast in OpenMPI in communication performance. Also, the evaluation implies that SDN-MPI-Bcast is feasible. © 2014 Springer-Verlag Berlin Heidelberg.

DOI： 10.1007/978-3-642-54420-0_86

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/europar/europar2013w.html#DashdavaaDYKWIAS13

Language：English   Publishing type：Research paper (scientific journal)   Publisher：Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)  

Collective communication is of great importance in MPI because the execution time of an MPI program is affected by the communication performance it can gain. Particularly these days, when a cluster system composed of multiple computing nodes has become dominant as a large-scale computing system, the execution time of collective communication affects the total execution time of the MPI program. However, in many implementations of MPI, collective communication is developed to make use of unicast-based communication in a repeated and combined way, which may result in inefficient communication. In this paper, we explore the use of a Software-Defined Network, which was originally expected to help network administrators operate networks through central control in a software-programming manner, to accelerate MPI-Bcast, a basic collective communication used in MPI. The evaluation in this paper indicates that our prototyped SDN-MPI-Bcast is superior to MPI-Bcast in OpenMPI in communication performance. Also, the evaluation implies that SDN-MPI-Bcast is feasible. © 2014 Springer-Verlag Berlin Heidelberg.

DOI： 10.1007/978-3-642-54420-0_86

Web of Science

Scopus

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

We have been involved to research and experiment of Long Distance Live Migration of virtual machines on wide area distributed storage which seems to be a single & shared large POSIX file system. In this paper, we introduce two key technologies in our project; "distributed storage" and "network". Distributed storage technology enables to achieve wide area distributed storage by distributing and sharing both meta data and user data in geographically distributed areas. We summarize network technologies in two parts, that is internal network and external network. Especially, the external network technology is important for Long Distance Live Migration so that external hosts (on the Internet, for example) would communicate with a virtual machine even the case where it would be migrated into geographically remote sites. In this paper, we also introduce the result of our experiment.

CiNii Books

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper (scientific journal)   Publisher：電子情報通信学会  

クラスタシステムなどの高性能計算環境では，複数のジョブ間でその資源が共有利用される．このような計算環境上で分散並列計算の実行時間を縮減するためには，その環境を構成する計算資源及びネットワーク資源を効率的に割り当てる必要がある．しかし，今日のジョブ管理システム（JMS）の多くは計算資源情報のみを考慮してジョブへの資源配分を決定する．本論文では，この点に着目し，ジョブごとに指定されるユーザの資源要求に基づき，計算資源及びネットワーク資源を考慮した資源割当を実現するためのジョブ管理フレームワークを提案する．技術的には，既存のJMSであるOpen Grid Scheduler/Grid Engine（OGS/GE）に，ネットワークの一元的な動的制御を可能にするOpenFlowを統合させることで実現したフレームワークの実装についてまとめる．本論文の評価では，提案するジョブ管理フレームワーク上に，ネットワークの遅延または帯域を考慮する資源割当ポリシを配備し，ネットワーク資源の割当をジョブごとに制御することでジョブの実行時間を縮減できることを確認した．

CiNii Books

J-GLOBAL

researchmap

Language：Japanese   Publisher：Information Processing Society of Japan (IPSJ)  

A large amount of scientific data needs to be transferred from one site to another as fast as possible in the computational science field. High-speed data transfer between sites is very important, especially in the Grid computing field; GridFTP has been widely used for bulk data transfer over wide area network. GridFTP achieves greater performance by supporting parallel TCP streams. Using parallel TCP streams improves the throughput of slow-start algorithm and lossy network even on a single path. This research proposes a traffic engineering technique that increases the data transfer performance by using multiple paths simultaneously for the parallel TCP streams. For the purpose, we use Software-Defined Network (SDN) technology and its implementation, OpenFlow. This paper presents the design and implementation of the proposed system and describes the results of the performance evaluation.

CiNii Books

J-GLOBAL

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

It is focused on distributed storage technologies and backing up data onto remote data centers because of growing sense of disaster recovery for large scale disaster. Now the authors are constructing widely distributed storage environment. This environment is characterized by high performance of random I/O access. Because every site can access to same data, IT services can be continued with live migration technology. In this paper, the authors explain the demonstrations on SC13 and SC14.

CiNii Books

J-GLOBAL

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

ICT systems are expected to ensure high availability against not only possible failures but also unexpected multiple failures due to wide-scale disaster. It is important to continuous evaluations on the operation through education or training in addition to verification on the design and development. In this paper, we propose a verification platform which enables to arbitrary-generate disasters or failures to evaluate disaster tolerance of the distributed system.

CiNii Books

J-GLOBAL

researchmap

Language：Japanese  

本稿では，複数大学のコンピュータリソースを相互に接続し，広域分散環境において頑強なデータ基盤を構築するためのアーキテクチャの提案を行う．本アーキテクチャは，多数のコンピュータリソースをもちいて大規模なストレージ空間を実現するクラスタストレージ技術を応用し，広域分散環境で信頼性の高いデータ基盤を実現する．本アーキテクチャによって実現されるデータ基盤は，地理的に離れた3カ所以上の「マルチサイト」から同時にアクセスが可能であり，かつ一カ所での更新がリアルタイムに他サイトに反映されることを特徴とする．同データ基盤は単一の巨大なファイルシステムとしてPOSIX準拠の標準的なアクセス手段を提供し，かつ，ランダムアクセスによるファイル入出力を可能とする．将来的には，地理的に離れた拠点間で仮想OSのライブマイグレーションができることを目指す．

CiNii Books

researchmap

Language：Japanese  

大規模災害による危機意識の高まりから災害回復（Disaster Recover: DR）を実現するための技術として遠隔地データセンターでのバックアップや分散ストレージに注目が集まっている．現在我々はランダムアクセス性能の高さに特徴のある広域分散ストレージ環境を金沢大学，広島大学，NIIを中心として構築しており，本研究では本環境のI/O性能を評価し，この環境の有用性を示す．

CiNii Books

researchmap

Language：Japanese  

J-GLOBAL

researchmap

Language：English   Publishing type：Research paper (scientific journal)   Publisher：Advances in Parallel Computing  

On 11 March 2011 JPN suffered a major earthquake. The resulting tsunami caused great loss of life and devastated many buildings, industries, and information services. Similarly, recent flooding in Thailand resulted in the destruction of property and a disruption of services. Thus, it is prudent for information technology infrastructure and services providers to formulate and implement procedures for rapid disaster recovery of key services that become even more critical during catastrophic events. To aid in our understanding of how academic resource centers can positively respond to such events, we have prototyped, developed and deployed GEO (Global Earth Observation) Grid applications on a distributed infrastructure made up of physical clusters contributed by our international colleagues. On top of these clusters, each site was able to host virtual machines, using their choice of virtualization infrastructure. We were able to realize a secure and transparent Inter-Cloud infrastructure through automated translation of virtual machines (e.g., between different hypervisors) and network virtualization. Cloud interoperation enables sharing of virtual machine images by private and public Clouds, i.e. a virtual machine image could be shared by different VM hosting environments including OpenNebula, Rocks, and Amazon EC2. Insights gained through the preliminary experiments indicated the key issue for Inter-Cloud is how we could use technologies for network virtualization. We used OpenFlow for network virtualization to build a secure (isolated) network for Inter-Cloud. © 2013 The Authors.

DOI： 10.3233/978-1-61499-322-3-35

Web of Science

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

The use of virtual cluster for High Performance Computing (HPC) work has a great benefit of hiding the complexity of physical infrastructure while providing a familiar HPC environment for users. Nevertheless, setting up one is still a challenging task for most users. In addition, the ability to span a virtual cluster across a multi-site cloud can even create a great potential for users to harness a much higher processing power from the cloud. In this paper, the development of a multi-site virtual cluster cloud is presented. This is done by providing the cloud infrastructures tailored specifically for virtual cluster deployment along with related utilities such as web-interface monitor and command line tools. To realize this infrastructure, an overlay network is created using Virtual Private Network (VPN) technology to merge every site into a single broadcast domain. Moreover, a Multi-tier VPN network topology is proposed to enhance the communication performance among the multi-site cloud. A smart virtual machine scheduling mechanism is introduced to reduce performance degradation caused by network delays. The experimental results show that the virtual cluster works well with highly parallel task with low communication among task. Thus, this technology can be used to support many classes of important HPC applications. © 2013 IEEE.

DOI： 10.1109/JCSSE.2013.6567337

Web of Science

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

Network performance in high-performance computing environments such as supercomputers and Grid systems takes a role of great importance in deciding the overall performance of computation. However, most Job Management Systems (JMSs) available today, which are responsible for managing multiple computing resources for distribution and balancing of a computational workload, do not consider network awareness for resource management and allocation. In this paper, the authors briefly overview our proposed and prototyped network-aware JMS that can allocate an appropriate set of computing and network resources to a job request. Also, we evaluate the usefulness and effectiveness of our proposal. Experiments conducted with the prototype implementation imply that our proposed network-aware JMS could reduce job execution time by 23.4 percent. © 2013 IEEE.

DOI： 10.1109/ICON.2013.6781934

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/icon/icon2013.html#WatashibaKDAIYKT13

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

Despite the development of virtualization technologies in computing resources, how best to build virtual networks is still an open issue. Virtual network is an important technology that can hide physical network topologies and policies and provide a dedicated virtual computing environment for each user. However, it is difficult to figure out the actual mapping between a physical network and a virtual network from a logical point of view; the difficulty will cause performance degradation in the virtual network and inefficient network routing. We have therefore proposed a network performance-aware routing mechanism using SDN technologies. By using SDN technology, we can control the network in a centralized manner. Our proposed system automatically aggregates the information on network performance, and then defines an appropriate routing policy by solving optimal path problems based on the observed network performance. We have improved the performance of the inter-domain network for a multisite virtual cluster with the proposed system. We have also started to deploy the proposed system into an actual environment consisting of different multiple organizations. © 2013 IEEE.

DOI： 10.1109/ICON.2013.6781935

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/icon/icon2013.html#IchikawaDAYKS13

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

Advances in sequencing technology have resulted in an exponential increase in the availability of protein sequence information. In order to fully utilize information, it is important to translate the primary sequences into high-resolution tertiary protein structures. MODELLER is a leading homology modeling method that produces high quality protein structures. In this study, the function of MODELLER was expanded by configuring and deploying it on a parallel grid computing platform using a custom four-step workflow. The workflow consisted of template selection through a protein BLAST algorithm, target-template protein sequence alignment, distribution of model generation jobs among the compute clusters, and final protein model optimization. To test the validity of this workflow, we used the Dual Specificity Phosphatase (DSP) protein family, which shares high homology among each other. Comparison of the DSP member SSH-2 with its model counterpart revealed a minimal 1.3% difference in output energy scores. Furthermore, the Dali Pairwise Comparison Program demonstrated a 98% match among amino acid features and a Z-score of 26.6 indicating very significant similarities between the model and actual protein structure. After confirming the accuracy of our workflow, we generated 23 previously unknown DSP family protein structure models. Over 40,000 models were generated 30 times faster than conventional computing. Virtual receptor-ligand screening results of modeled protein DSP21 were compared with two known structures that had either higher or lower structural homology to DSP21. There was a significant difference (p≤0.001) between the average ligand ranking discrepancy of a more homologous protein pair and a less homologous protein pair, suggesting that the protein models generated were sufficiently accurate for virtual screening. These results demonstrate the accuracy and usability of a grid-enabled MODELLER program and the increased efficiency of processing protein structure models. This workflow will help increase the speed of future drug development pipelines. Copyright © 2013 by The Institute of Electrical and Electronics Engineers, Inc.

DOI： 10.1109/eScience.2013.15

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/eScience/eScience2013.html#LiTXHID13

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

Whether to be able to efficiently utilize the interconnect in a cluster system is a key factor in deciding the computational performance especially for a class of jobs that require intensive communications between processes. However, most of Job Management Systems (JMSs), which are deployed on cluster systems for load-balancing, do not have any mechanism that takes the pattern and requirements of communication occurred in a job into consideration, regardless of potential large impact on its performance. In this research, we explore a novel JMS that can assign computing resources to jobs submitted to the JMS from a standpoint of efficient use of both network and processor resource. More technically, in this paper, a JMS that integrates the network programming functionality of OpenFlow as Software Defined Network is proposed and discussed towards better and more efficient allocation of computing and network resources. © 2013 IEEE.

DOI： 10.1109/COMPSACW.2013.108

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/compsac/compsac2013w.html#WatashibaDAIYKT13

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

Recently, the concept of Software-Defined Network (SDN), which allows us to administer and configure a network in a centralized and software-programming manner, has gathered network engineers' and researchers' attention rapidly. In particular, the expectation and concern to OpenFlow as an implementation of the SDN is remarkable. As a result, research activities, which include prototyping, implementation, demonstration and experiments, conducted over OpenFlow networks have been a worldwide tendency. In such research activities, however, the difficulty in understanding network topology, traffic amount and an actual path of a network flow on the OpenFlow network, and the intricacies in debugging software designed for OpenFlow are serious problems in the development process of OpenFlow controller. This research aims to realize a visualization software that facilitates researchers to perform OpenFlow controller development and demonstration experiments performed on an actual OpenFlow network. In this paper, the authors summarize the achievement of their research work in progress as well as the future direction. © 2013 IEEE.

DOI： 10.1109/COMPSAC.2013.81

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/compsac/compsac2013.html#WatashibaHDAIKST13

Publishing type：Research paper (scientific journal)  

DOI： 10.1007/s12626-013-0038-0

researchmap

Language：Japanese   Publisher：(公社)全国大学保健管理協会  

本学(大学)保健センターでは、本年度(2012年度)より学生定期健康診断(学生健診)にWEB調査票システムを導入しており、今回、その効果について検討した。学生健診をWEB化したことで、導入前に予想していたペーパレス化による混雑緩和、部局ごとの回答率把握といったメリットが得られたが、混雑予測の把握のための混雑予測カレンダーが有効に機能していなかったことなどが、今後の改善点として明らかになった。WEB化したことによる受検率の変化はほとんどみられず、72.7%で前年度比約370名の減少であった。

J-GLOBAL

researchmap

Language：Japanese  

近年，高性能計算資源はクラスタシステムとして構成されていることが主流で，より高性能を求めてますます大規模化する傾向にある．複数の計算ノードがインターコネクトによって相互接続されたクラスタシステムでは，ユーザがより高い実行性能を得るには，できるだけ多数の計算資源を高効率に利用する分散並列処理を行う必要がある．とりわけ，計算ノード間の通信が密な計算では，割り当てられた計算ノード間のネットワーク性能が計算性能に多大な影響を与える．しかし，今日利用できる，計算ノードを管理しユーザジョブの計算要求に応じて，効率的に計算ノードの計算資源を割り当てる役割を担うジョブ管理システムの多くは，計算ノード間のネットワーク，すなわちインターコネクトの効率的利用を考慮した仕組みを有していない．われわれは，そのような視点から，ネットワーク資源も含めた計算資源の適切な管理・割当を行うことができるジョブ管理システムの実現を目的とした研究開発に着手した．本稿では，計算ノード間のインターコネクトを Software-Defined Network (SDN) で制御することにより，計算性能の向上を目的とするジョブ管理システムを提案し，提案システムの有用性についてケーススタディを用いて検討する．High performance computing has been becoming more large-scale and distributed. Today, since the current dominant trend of high-performance computing is cluster system, it tends to be built with more computers for more performance improvement and acceleration. Therefore, the performance of interconnect among computing resources of the cluster plays more important role due to a fact that parallel computation performed on distributed multiple resources requires communication. However, no current Job Management System (JMS) is designed to consider how network resources, that is, interconnects are used in submitting jobs to computing resources of a cluster system. We have started the research and development of network-aware job management system that allocates appropriate set of computing and network resources in an on-demand fashion, based on user-provided resource requirement information. The key feature of our proposed JMS is the integration of Software-Defined Network (SDN) concept into traditional JMS.

CiNii Books

J-GLOBAL

researchmap

Language：Japanese   Publisher：一般社団法人電子情報通信学会  

本稿では,複数大学のコンピュータリソースを相互に接続し,広域分散環境において頑強なデータ基盤を構築するためのアーキテクチャの提案を行う.本アーキテクチャは,多数のコンピュータリソースをもちいて大規模なストレージ空間を実現するクラスタストレージ技術を応用し,広域分散環境で信頼性の高いデータ基盤を実現する.本アーキテクチャによって実現されるデータ基盤は,地理的に離れた3ヵ所以上の「マルチサイト」から同時にアクセスが可能であり,かつ一ヵ所での更新がリアルタイムに他サイトに反映されることを特徴とする.同データ基盤は単一の巨大なファイルシステムとしてPOSIX準拠の標準的なアクセス手段を提供し,かつ,ランダムアクセスによるファイル入出力を可能とする.将来的には,地理的に離れた拠点間で仮想OSのライブマイグレーションができることを目指す.

CiNii Books

J-GLOBAL

researchmap

Language：Japanese   Publisher：一般社団法人情報処理学会  

大規模化・分散化傾向にあるクラスタ型計算資源の利用において,ネットワーク性能が計算処理性能に与える影響は大きい.しかし,ユーザジョブの計算要求に応じて計算資源を割り当てるジョブ管理システムの多くは,ネットワーク資源情報を利用せず,CPU やメモリのような計算資源情報に基づいて各ジョブへの割当資源を決定する.本研究ではネットワーク資源と計算資源の情報に基づき,ユーザのジョブ要求に対して最適なネットワークを持つ計算資源を割り当てることができるジョブ管理システムの実現を目指している.本稿では,ネットワークを柔軟に制御することが可能な技術として注目を集めている OpenFlow の機能を応用したジョブ管理システムの概要をまとめるとともに,提案システムの有効性を検証した実験結果について報告する.

CiNii Books

J-GLOBAL

researchmap

Language：Japanese   Publisher：一般社団法人情報処理学会  

科学全般において大規模な可視化が必用とされ,可視化装置とミドルウェアに注目を集めている.複数のディスプレイで構築するタイルドディスプレイは大規模な可視化を実現する技術であり,ネットワークを通じてアプリケーションやコンテンツを表示することができる.動画配信などの大容量ストリーミングとは異なり,タイルドディスプレイでのアプリケーション表示ではユーザのインタラクションが発生し,その結果を画面に反映することとなるが,ネットワークの遅延,障害がユーザインタラクションに多大な影響をおよぼす.そのためタイルドディスプレイでのユーザインタラクションを考慮したネットワークの制御が必要不可欠となる.先行研究では,SAGE によるタイルドディスプレイと OpenFlow を用い障害検知による冗長経路の変更について取り組んだ.本研究では,ユーザ操作によって生じるデータ転送量の変更に対応するため OpenFlow を用いたフロー制御に取り組み,インタラクティブネスの向上を目指す.

CiNii Books

J-GLOBAL

researchmap

J-GLOBAL

researchmap

Language：Japanese   Publisher：一般社団法人電子情報通信学会  

大規模災害による危機意識の高まりから災害回復(Disaster Recover:DR)を実現するための技術として遠隔地データセンターでのバックアップや分散ストレージに注目が集まっている.現在我々はランダムアクセス性能の高さに特徴のある広域分散ストレージ環境を金沢大学,広島大学,NIIを中心として構築しており,本研究では本環境のI/O性能を評価し,この環境の有用性を示す.

CiNii Books

J-GLOBAL

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

Transferring virtual clusters among data centers has received much attention recently. However, since each VM disk image size is huge and a virtual cluster is composed of multiple VMs, transferring a virtual cluster takes a long time. To address this issue, we have proposed multiple disk image deduplicating methods that take the physical deployment structure of VMs into consideration. This paper evaluates and compares multiple deduplication methods, and clarifies the characteristics of each method. © 2013 International Information Institute.

Scopus

CiNii Books

researchmap

Language：Japanese   Publishing type：Research paper (scientific journal)   Publisher：情報処理学会  

Recently, wide-area distributed computing has been increasingly concerned. Particularly in the scientific research, virtual cluster, which aggregates virtual machines from multiple physical computers to provide a user-dedicated computing environment, has attracted scientists and researchers expectation. In reality, however, no practically available virtual cluster solution has been realized so far. This fact is explained from the difficulty of establishing a network where virtual machines composing a virtual cluster directly communicate and that of installing and configuring software stack as a single cluster system on them. In this paper, we have proposed and prototyped a multi-site virtual cluster deployment system that builds a virtual cluster aggregating virtual machines, each of which is located on a different site. To evaluate the practicality, we have measured the computational performance of a virtual cluster system deployed on a WAN-emulated environment. The result indicates that the virtual cluster provides computational performance enough to perform a parameter-study problem with less data transfer.

CiNii Books

researchmap

Other Link： http://id.nii.ac.jp/1001/00086045/

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE Computer Society  

A tiled display wall (TDW), which is a single large display device composed of multiple sets of computers and displays, has recently gained the attention of scientists. In particular, SAGE, which is a middleware for building TDW, allows scientists to browse a multiple series of visualized results of computer simulation and analysis through the use of network streaming. Each visualized result can be generated on a different remote computer. For this reason SAGE has been increasingly hailed as a promising visualization technology that will solve the geographical distribution problem of computational and data resources. SAGE depends heavily on a network streaming technique in its architecture, but does not have any recovery mechanism against cases of network problems. This research, therefore, aims at realizing a network failure avoidance functionality for SAGE, focusing on OpenFlow against SAGE vulnerability to network failure. Specifically, the functionality is designed and developed as a composition of the following three functions: network failure detection, network topology understanding, and a packet forwarding control function. The key concept behind our design is that the network control function from OpenFlow should be built into SAGE. The evaluation in the paper confirms that the proposed and prototyped network failure avoidance functionality can detect failures on network routes and then reroute network streaming for visualization on TDW. © 2012 IEEE.

DOI： 10.1109/COMPSACW.2012.26

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/compsac/compsac2012w.html#FuruichiDYIATK12

Language：Japanese   Publisher：情報処理学会  

データセンタの需要増に伴い、データセンタ内におけるネットワーク機器の消費電力が増加している。OpenFlow技術を用いて、電力量を少なくするパスにトラフィックを集約し、不要なスイッチを休止させる手法を提案する。

J-GLOBAL

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Datacenter is essential for service provider on the Internet. However, energy consumption of ICT equipment is increasing each year. It is important for datacenter administrator to reduce energy consumption of datacenter which contains a large amount of ICT equipment. Improving server machine or cooling facilities of efficiency, network resource consumes more energy. In this paper, We prototyped event-driven OpenFlow controller which finds minumum-power network subsets using Best-fit algorithm and flow contention algorithm. Evaluation results shows that prototype can reduce datacenter network energy consumption.

CiNii Books

J-GLOBAL

researchmap

Language：Japanese   Publisher：(公社)全国大学保健管理協会  

J-GLOBAL

researchmap

Language：Japanese  

近年，仮想計算機 (VM: Virtual Machine) 技術の発達に伴い，仮想クラスタを異なる拠点間で移送する拠点間移送に関心が高まりつつある．しかし，仮想クラスタの移送では，データサイズの大きい VM ディスクイメージを多数転送する必要があるため，移送時間が長くなることが問題である．そこで，本研究においては仮想クラスタを構成する複数の VM ディスクイメージ間には重複内容が多いことに着目し，これらの重複内容の除去を行うことにより，移送時間の削減を図った．具体的には，重複除去の手法として，ローカル重複除去，多段階重複除去の2種類の重複除去手法を提案し，これらを実行するためのプログラムの実装を行った．本研究では，これらの実装したプログラムの処理時間の測定を行い，VM のサイズや個数，使用可能帯域幅などの任意の条件下における仮想クラスタ移送時間の推定式を導出した．この推定式を利用することにより，WAN 実効帯域幅等の条件に応じて，2種類の重複除去手法，および重複除去を行わずに移送する手法の3種類のうち，いずれが移送時間を最短にするものであるかの決定指針を与えることができた．Transferring virtual clusters across different sites has been gathering much attention recently due to the development of virtual machine technology. However, it takes a long time to transfer VM disk images composing a virtual cluster across different sites since the total size of multiple VM disk images is very large. In this study, we proposed an efficient transferring method that reduces the size of transferring data by deduplicating multiple VM disk images composing a virtual cluster. In particular, we have developed two deduplication methods, local deduplication and multistage deduplication. We have evaluated the execution time of the developed duduplication programs, and derived estimating equations of transferring time of virtual cluster under arbitrary conditions.

J-GLOBAL

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IOS PRESS  

On 11 March 2011 JPN suffered a major earthquake. The resulting tsunami caused great loss of life and devastated many buildings, industries, and information services. Similarly, recent flooding in Thailand resulted in the destruction of property and a disruption of services. Thus, it is prudent for information technology infrastructure and services providers to formulate and implement procedures for rapid disaster recovery of key services that become even more critical during catastrophic events. To aid in our understanding of how academic resource centers can positively respond to such events, we have prototyped, developed and deployed GEO (Global Earth Observation) Grid applications on a distributed infrastructure made up of physical clusters contributed by our international colleagues. On top of these clusters, each site was able to host virtual machines, using their choice of virtualization infrastructure. We were able to realize a secure and transparent Inter-Cloud infrastructure through automated translation of virtual machines (e.g., between different hypervisors) and network virtualization. Cloud interoperation enables sharing of virtual machine images by private and public Clouds, i.e. a virtual machine image could be shared by different VM hosting environments including OpenNebula, Rocks, and Amazon EC2. Insights gained through the preliminary experiments indicated the key issue for Inter-Cloud is how we could use technologies for network virtualization. We used OpenFlow for network virtualization to build a secure (isolated) network for Inter-Cloud. © 2013 The Authors.

DOI： 10.3233/978-1-61499-322-3-35

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/hpc/hpc2012-1.html#TanakaYTOPWZHPW12

Publishing type：Research paper (scientific journal)   Publisher：Korea Information Processing Society  

DOI： 10.3745/jips.2011.7.4.581

researchmap

Language：English  

researchmap

Language：English  

researchmap

Language：Japanese   Publishing type：Research paper (scientific journal)   Publisher：The Japanese Society for Artificial Intelligence  

DOI： 10.11517/jjsai.26.2_117

CiNii Books

J-GLOBAL

researchmap

Other Link： http://id.nii.ac.jp/1004/00007736/

Language：Japanese   Publisher：情報処理学会  

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE Computer Society  

This paper introduces an application of the data mining tool, MUSASHI (Mining Utilities and System Architecture for Scalable processing of Historical data), for scientific policy making. Recent advances in information systems have allowed researchers to gather enormous amounts of data on opinions about government policy. However, these gathered data have been individually stored, and have never been integrated because of a lack of techniques to analyze the data in an integrated way. In this paper we propose a cloud application for policymaking during a financial crisis. For this purpose, we have developed an ASP- platform- leveraging distributed- computing technology represented by Cloud computing. © 2011 IEEE.

DOI： 10.1109/GRC.2011.6122693

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/grc/grc2011.html#YadaI11

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE Computer Society  

Visualization has been taking a role of much importance in both science and business areas. For scientists, visualization is a method to gain the meaning of and insight from scientific data and computational results; for business marketing area, computer visualization is popular as a digital signage technology. Recently, tiled display wall (TDW) has gathered users' attention and concern as a promising large-scale visualization technique. SAGE (Scalable Adaptive Graphics Environment) is a representative middleware for realizing such a TDW and experimentally used in an inquisitive way at research institutions and universities. However, this SAGE has an undesirable problem that existent X-window based visualization applications are not displayed without source code modification, although it provides a suite of APIs with which developers can write visualization applications displayed on the SAGE-enabled TDW. In this paper, we present an adapter solution that allows users to display X-window application on the SAGE-based TDW without source code modification. In our evaluation, the developed adapter solution incurred a small overhead in displaying an X application onto a SAGE-enabled TDW, in comparison with the case of modifying the source code of it. © 2011 IEEE.

DOI： 10.1109/GRC.2011.6122667

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/grc/grc2011.html#TadaDSIA11

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE Computer Society  

Providing large scale, stable and robust computing environments is indispensable to various fields including scientific researches as well as business. Grid computing has gathered considerable attention as a means to efficiently and effectively utilize computational resources among multiple organizations. However, the current Grid technologies do not always provide the computational environment enough to satisfy users' requirements because of the heterogeneity and disconnect among computational resources on the Grid. For the reason, we have been attempting to develop a new middleware that leverages newly emerging Cloud technologies and allows for a virtual cluster composed of virtual computational resources by Xen as a virtualization technology on top of an overlay network realized by PIAX as an overlay networking technology. In this paper, we propose a system that allows Xen-based virtual computational resources composing a virtual cluster to communicate with each other toward the future development of Grid and Cloud middleware technology. The evaluation shown in the paper indicates that such a virtual cluster is feasible and could work effectively for computationally intensive applications. © 2011 IEEE.

DOI： 10.1109/GRC.2011.6122611

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/grc/grc2011.html#IchikawaDKS11

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1007/s12626-010-0016-8

researchmap

Language：Japanese  

近年，大規模データ可視化技術として，タイルドディスプレイが注目されており，その実現技術として SAGE と呼ばれるミドルウェアがスケーラビリティやパフォーマンスの高さから科学研究への応用に期待されている．しかし，SAGE は既存の可視化アプリケーションをそのままでは利用できないという相互運用性に問題がある．本研究では，この問題に対して，可視化アプリケーションを変更することなく，アプリケーションから可視化データを取得し，SAGE によるタイルドディスプレイに表示可能とする手法を提案し，それを実現する SAGE アダプタの実装を行った．本稿では，この SAGE アダプタについて詳しく述べると共に，既存のアプリケーションを用いた動作検証結果を示す．また，実用上十分な性能が得られることを評価結果と共に報告する．Tiled Display has recently gathered wide attention as a possible way of realizing large-scale visualization. In particular, SAGE is a representative middleware to build up such a tiled display; it is highly expected to be utilized for scientific researches due to its scalable and high performance nature. However, SAGE has an interoperability problem that existing visualization applications are not available without modification of the implementation. To tackle this problem, we proposed SAGE adapter, which allows users to utilize the existing visualization applications in SAGE without modification. This paper describes the details of the SAGE adapter, and introduces an example of an application using SAGE adapter. Moreover the paper reports some evaluation results, which indicate that the SAGE adapter has enough performance for practical use.

CiNii Books

J-GLOBAL

researchmap

Language：Japanese  

近年，ネットワークを利用するサービスに対する攻撃が増加および多様化しており，NIDS (Network-based Intrusion Detection System) の重要性が高まっている．今日の NIDS にはネットワークの広帯域化に応じた高速な処理が求められており，高性能な NIDS の実現が急務である．このような背景から，本論文ではロードバランサを用いて NIDS の処理を並列化する並列 NIDS の実現を目的とし，高性能なソフトウェアロードバランサを設計，実装する．提案実装では高速なパケットキャプチャおよび転送モジュールである PF RING を利用し，パケット転送処理をマルチスレッドで並列化することで高速なパケット転送を実現する．さらに，本論文の性能評価によって，提案実装におけるパケット転送が 9.8Gbps を超えるスループットを達成し，既存研究のソフトウェアロードバランサで採用されているパケット転送よりも高い性能を持つことを示す．With the increase and diversity of attacks against network services, NIDS (Network-based Intrusion Detection System) has been taking a role of importance. For the traffic analysis of the today&#039;s high-speed network, the demand on high performance NIDS has been increased. From the background, we focus on parallel distributed NIDS that analyzes network traffic on multiple PCs in a high-throughput fashion. In this paper, we design and implement a high performance software-based load balancer that distributes the workload of traffic analysis to multiple commodity PCs. A prototypic implementation of the proposed software-based load balancer has realized high-throughput packet forwarding by multi-threaded processing and PF RING which provides high-speed packet capturing and sending. Our experiments indicate that, packet forwarding in the proposed software-based load balancer has achieved 9.8 Gbps throughput.

CiNii Books

J-GLOBAL

researchmap

Language：Japanese  

CiNii Books

J-GLOBAL

researchmap

Language：English  

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE Computer Society  

The objective of this paper is to introduce further development of a data mining tool, MUSASHI (Mining Utilities and System Architecture for Scalable processing of HIstorical data), for service computing. Recent advances in information systems have allowed us to gather enormous amounts of data on marketing. However, these gathered data have been individually stored at each company, and have never been integrated because of a lack of techniques to analyze the data in an integrated way and to handle the large amount of data efficiently. To address this issue, we are currently investigating a way to provide a data mining platform as a service so that users can apply various data mining techniques to their marketing data with ease and at a low cost. For this purpose, we have developed an ASP platform leveraging distributed computing technology rep- resented by Cloud computing. This paper describes the ASP platform for data mining services and introduces an empirical application of data mining using our platform. © 2010 IEEE.

DOI： 10.1109/GrC.2010.168

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/grc/grc2010.html#IchikawaYW10

Publishing type：Research paper (scientific journal)   Publisher：Japan Society for Software Science and Technology  

SAGE has received attention as a middleware that allows scientists who use large-scale data to control and drive a Tiled Display, which provides high-quality and high-density visualization with multiple displays. However, SAGE has a problem of poor application operability. This problem is due to the architectural design of SAGE that users have to manage window-related operation and application-related operation from different interfaces. To address this problem, we have developed an application event control module for SAGE. This module unifies the two separated interfaces of window management and application operation. Our design principle is to minimize the current SAGE architecture changes and keep portability of the module. In this paper, we describe the function, design and implementation of our module. To evaluate interactivity and operability of our module, we measured time taken from coming up an event to the completion of event reaching an application and time taken for users to perform pointing operation through SAGE UI. As a result, we confirmed that our module provides realistic interactivity and allows users to smoothly perform pointing operations.

DOI： 10.11309/jssst.27.4_172

Scopus

J-GLOBAL

researchmap

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

Recently, both networking and high-performance computing technologies have dramatically advanced. As a result, scientists can access large-scale data captured by scientific measurement devices with ease. Also, recent high-performance computing technologies such as the Grid and cloud computing have allowed scientists to use huge computational power harnessed on the Internet for their analysis of scientific data. However, visualization technology, which plays the role of great importance for scientists intuitively in understanding the analysis results of such scientific data, is not fully utilized so that it can benefit from recent high-performance and networking technologies although many sophisticated visualization technologies exists nowadays. One of such visualization technologies is SAGE (Scalable Adaptive Graphics Environment), which allows people to build an arbitrary size of a tiled display wall and is expected to be applied to scientific research. In this paper, we describe a multi-application controller for SAGE middleware, which we have developed, in the hope that it will help scientists perform their analysis on a wide-area distributed computing environment. The evaluation in the paper indicates the efficiency of completing a comparison task among multiple data is raised by our system. © 2010 IEEE.

DOI： 10.1109/HUMANCOM.2010.5563341

Scopus

researchmap

Language：Japanese   Publishing type：Research paper (international conference proceedings)  

Resource discovery is a fundamental service of the Grid to find out resources suitable for executing user obs. Traditional resource discovery services of the Grid have static structure and centralized or hierarchized, such as Globus MDS. However, with the number of resources composing the Grid increased, this approach has two problems, namely, on dynamic changes of the structure in the Grid and obtaining the latest resource information. To solve these problems, we have been seeking for the possibility of a resource discovery mechanism adopting Peer-to-Peer approach, which is characterized by real-time retrieval of information using overlay networks spanned on physical computational resources. In this paper, we present the preliminary results of our research activities so far. © 2009 IEEE.

DOI： 10.1109/SAINT.2009.68

Scopus

CiNii Books

researchmap

Language：English  

researchmap

Authorship：Lead author   Language：English   Publishing type：Research paper (international conference proceedings)  

researchmap

Language：English   Publishing type：Research paper (scientific journal)   Publisher：ELSEVIER IRELAND LTD  

Grid computing offers the powerful alternative of sharing resources on a worldwide scale, across different institutions to run computationally intensive, scientific applications without the need for a centralized supercomputer. Much effort has been put into development of software that deploys legacy applications on a grid-based infrastructure and efficiently uses available resources. One field that can benefit greatly from the use of grid resources is that of drug discovery since molecular docking simulations are an integral part of the discovery process. In this paper, we present a scalable, reusable platform to choreograph large virtual screening experiments over a computational grid using the molecular docking simulation software DOCK. Software components are applied on multiple levels to create automated workflows consisting of input data delivery, job scheduling, status query, and collection of output to be displayed in a manageable fashion for further analysis. This was achieved using Opal OP to wrap the DOCK application as a grid service and PERL for data manipulation purposes, alleviating the requirement for extensive knowledge of grid infrastructure. With the platform in place, a screening of the ZINC 2,066,906 compound "drug-like" subset database against an enzyme's catalytic site was successfully performed using the MPI version of DOCK 5.4 on the PRAGMA grid testbed. The screening required 11.56 days laboratory time and utilized 200 processors over 7 clusters.

DOI： 10.1016/j.cmpb.2008.07.005

Web of Science

Scopus

PubMed

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：SPRINGER-VERLAG BERLIN  

This research aims to present an analysis to optimally allocate advertising budgets based on single source data on consumers' views of TV advertising. A model of consumer behavior and an optimality criterion for the advertising budget allocation are proposed together with a GA based optimization algorithm. Through the analysis, we discovered some knowledge to improve the effectiveness of advertising for several products. © 2009 Springer Berlin Heidelberg.

DOI： 10.1007/978-3-642-04592-9_34

Web of Science

Scopus

researchmap

Language：Japanese   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

Resource discovery is a fundamental service of the Grid to find out resources suitable for executing user obs. Traditional resource discovery services of the Grid have static structure and centralized or hierarchized, such as Globus MDS. However, with the number of resources composing the Grid increased, this approach has two problems, namely, on dynamic changes of the structure in the Grid and obtaining the latest resource information. To solve these problems, we have been seeking for the possibility of a resource discovery mechanism adopting Peer-to-Peer approach, which is characterized by real-time retrieval of information using overlay networks spanned on physical computational resources. In this paper, we present the preliminary results of our research activities so far. © 2009 IEEE.

DOI： 10.1109/SAINT.2009.68

Web of Science

Scopus

CiNii Books

researchmap

Publishing type：Research paper (international conference proceedings)   Publisher：IEEE Computer Society  

DOI： 10.1109/SAINT.2009.68

J-GLOBAL

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/saint/saint2009.html#KokuboIID09

Authorship：Lead author,　Corresponding author   Language：English   Publishing type：Research paper (international conference proceedings)  

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

The Grid allows us to integrate computational resources widely distributed on the Internet to a large-scale computational environment. However, it is still hard to execute applications such computational environment as flexibly as a single cluster system because of the heterogeneity of software components of computational resources composing the Grid and the difficulty of establishing bi-directional connection among computational resources. Motivated by these reasons, we have been working on the development of the technologies which provide users with user-optimal flexible computational environment by combining virtual machine technologies and overlay networking. In this paper, our envisaged virtual cluster is overviewed together with its technical issues. After that, the communication relay mechanism using overlay networking is detailed. We summarize the results of performance test performed on a virtual cluster using the developed communication relay mechanism.

CiNii Books

J-GLOBAL

researchmap

Publishing type：Research paper (international conference proceedings)   Publisher：IEEE Computer Society  

Recently, Grid computing has gathered the attention as a means to efficiently and effectively utilize computational resources among multiple organizations. However, the current Grid technologies do not always provide the computational environment enough to satisfy users' requirements because of the heterogeneity and disconnect among computational resources on the Grid. For the reason, we have been attempting to develop a new Grid middleware technology that allows for a virtual cluster composed of virtual computational resources by Xen as a virtualization technology on top of an overlay network realized by PIAX as an overlay networking technology. In this paper, we propose a communication relay mechanism that allows Xen-based virtual computational resources composing a virtual cluster to communicate with each other toward the future development of such a Grid middleware technology. The evaluation shown in the paper indicates that such a virtual cluster is feasible and could work effectively for computationallyintensive applications. © 2008 IEEE.

DOI： 10.1109/SAINT.2008.94

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/saint/saint2008.html#KusumotoDITS08

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

Virtual screening has become an important part of the drug discovery process. Grid computing facilitates this process by providing shared computational resources across different international institutions to run computationally intensive, scientific applications without the need for a centralized supercomputer. This study designed and implemented a flexible, scalable platform to perform a large virtual screening experiment on the PRAGMA grid testbed using the molecular docking simulation software DOCK 5.4. Using Opal OP to wrap DOCK as a grid service and PERL for data manipulation purposes, the "druglike" subset of the ZINC database, which contains 2,066,906 compounds, was successfully screened against the catalytic site of a protein tyrosine phosphatase. The screening required 11.56 days laboratory time and utilized 200 processors over 7 clusters. A ranked list of the best binding compounds to the phosphatase was generated and is currently being tested in biological applications for their efficacy and specificity.

DOI： 10.1109/GRID.2008.4662800

Web of Science

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/conf/grid/2008

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE Computer Society  

In 2005, Osaka University, in Japan, started an international educational program called, Pacific Rim International University (PRIUS), on top of the Pacific Rim Application and Grid Middleware Assembly (PRAGMA) research framework. The PRIUS framework is based on and similar to that of the PRIME program at the University of California San Diego. Through the PRIUS program, Osaka University has explored a new structure of higher educationfor graduate students by combining lectures given by PRAGMA researchers and scientists as well as internship abroad opportunities to PRAGMA member institutions and universities. In this paper, we describe the goals and framework of the PRIUS program and discuss issues for the improvement of PRIUS. We also present two examples of interns' achievements as well as other educational effects brought through collaboration with PRAGMA. © 2008 IEEE.

DOI： 10.1109/eScience.2008.84

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/eScience/eScience2008.html#DateMISTF08

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE Computer Society  

SHP-2 is a protein tyrosine phosphatase (PTP) that plays an important role in many cellular functions such as development, growth, and death; thus SHP-2 has been hypothesized to play an important role in various diseases such as diabetes, neurodegeneration, and cancer. The importance of the individual roles of different PTPs is not well understood and this is complicated by the lack of specific inhibitors. In this study, we have utilized the multi-institutional PRAGMA Grid computation resources to virtually screen the ZINC 7 database using virtual docking software DOCK 6.2. Preliminary results suggest several SHP-2 specific inhibitors that can be further tested and validated under laboratory conditions. Complications during these multiple, virtual screenings on the grid as well as potential improvements are also discussed. These findings have future clinical significance in the creation of new drug therapies for the treatment of different diseases.

DOI： 10.1109/eScience.2008.40

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/eScience/eScience2008.html#HanLIDH08

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE Computer Society  

The slingshot-2 (SSH-2) protein plays a significant role in different cell functions such as growth and movement. SSH-2 is a phosphatase protein that belongs to a unique class of enzymes called dual specificity phosphatases (DSP) that target the phosphothreonine and phosphotyrosine residues of mitogen-activated protein (MAP) kinases, which regulate cell growth. Because of this, it is of great interest to find specific inhibitors of DSPs such as SSH-2. Implementing an in silico platform to screen a sizable pool of chemical compounds against SSH-2 on the grid environment with the molecular docking software DOCK 6, several chemical compounds have been identified as potential inhibitors of SSH-2 activity. The issues of performing routine virtual screenings on the grid and possible improvements are also presented. The most promising inhibitor determined from standard and AMBER DOCK screenings was 2-amino-3-phosphonooxy-propanoic acid, which will be verified with wet bench testing.

DOI： 10.1109/eScience.2008.39

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/eScience/eScience2008.html#PhamLIDH08

Language：English   Publishing type：Research paper (other academic)  

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：WORLD SCIENTIFIC PUBL CO PTE LTD  

This paper describes a new approach (Opal OP: Opal Operation Provider) to wrap existing legacy applications as Grid services. In order to expose, with minimal effort, existing applications as Grid services, Opal OP provides a. method for wrapping a legacy application as a. program module, or as an operation provider. Traditional wrapping methods usually restrict the way to implement Grid services because these methods provide only a suite of interfaces necessary for using the wrapped application. The proposed Opal OP, on the other hand, doesn't restrict the way to implement; a Grid service from a legacy application. Opal OP is implemented its an Operation provider that wraps the application and thus can be used is a module in the Grid service. Application developers Call easily develop their own services where legacy applications are wrapped through the utilization of Opal OP. In this paper, we show some scientific applications including a bio-molecular simulation system developed as Grid services using Opal OP. The results show the usefulness and effectiveness of Opal OP.

DOI： 10.1142/9789812708823_0010

Web of Science

researchmap

Language：Japanese   Publishing type：Research paper (other academic)  

J-GLOBAL

researchmap

Language：Japanese   Publisher：Information Processing Society of Japan (IPSJ)  

SAGE has been gathering researchers' attention as a visualization technology that enables enormous scientific data in a high-resolution manner. This SAGE is the technology that displays enormous scientific data on a local high-resolution display device through the network streaming of the scientific data from a remote site. Today, this characteristic of SAGE is making the development of collaboration environment where researchers who perform their own scientific research through the sharing and visualization of scientific data active. In this paper, we first summarize and briefly explain these technical issues around SAGE and then clarify the research issues from now on. Subsequently, the interface problem related to window and application operations, which is one of such technical issues, is mentioned and its solution is shown in this paper.

CiNii Books

J-GLOBAL

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

Recently, the amount of scientific data has been increasing because of technological advancement in measurement and high-performance computing. As the result, visualization technologies are becoming increasingly important to analyze the scientific data. Among visualization technologies, SAGE is gathering attention of scientists in these days. It not only enables the visualization of tens of million pixel-class high-resolution image but also has the potential to realize the collaboration environment where scientists can work together for their common research purpose. However, there remain still technical issues in SAGE, since it is still in the process of development. A problem of application operability, which prevents SAGE users from efficiently operating various applications on SAGE, is targeted in this paper. This problem is caused from the architectural design that SAGE users have to manage window-related operation and application-related operation from different interfaces. For solving the problem, we have developed a built-in application control module for SAGE. Our built-in application control module unifies the separate interfaces for window management and application operation. As a result, users can control application from a unified user interface, SAGE UI. ©2007 IEEE.

DOI： 10.1109/PACRIM.2007.4313191

Web of Science

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

The maturity and stability of WSRF and its related technologies is producing a new demand for meta-scheduling WSRF-based services. In fact, many scientific institutions and universities are looking into the possibility of using WSRF-based services as production-level services. For this purpose, they need a meta-scheduling solution that selects one of the WSRF-based services, each of which is deployed in an organization for fault tolerance and load balancing. However, methods for meta-scheduling services among multiple sites have not been discussed well. This paper describes a new framework (MSSA: Meta-Scheduling Service Architecture) for scheduling and selecting WSRF-based services deployed over multiple sites. This framework provides a transparent scheduling method to users. Our method takes advantage of the factory pattern used typical WSRF services. The framework builds scheduling processes into the factory pattern and hides the complexity of the scheduling process from users. MSSA defines a meta-factory service that selects a suitable factory service and delegates the user's request to the factory service. By using this meta-factory service instead of using the original factory services, users can utilize meta-scheduling features transparently. MSSA can build hierarchical structures of meta-factory services to construct scalable scheduling environment, and can also switch scheduling policies easily. In this paper, the design and implementation of MSSA is discussed. ©2007 IEEE.

DOI： 10.1109/PACRIM.2007.4313278

Web of Science

Scopus

researchmap

Language：Japanese   Publisher：Information Processing Society of Japan (IPSJ)  

The development of High-performance computing technology has led to the advancement in computer simulation technology. In the arena of life science, experts have been attempting the combinational and integrated use of such computer simulation programs developed until today with a new perspective to the reveal of biological system. In this paper, we describe a simulation system for understanding biological system, characterized by the use of OPAL. The OPAL allows us to easily make the existent simulation programs a web service in a wrapping manner. Also, OPAL Operation Provider that allows the web service realized with OPAL to utilize Grid function is detailed. After that, we consider the usefulness of the system taking advantage of OPAL.

CiNii Books

J-GLOBAL

researchmap

Publishing type：Research paper (international conference proceedings)   Publisher：IEEE Computer Society  

Cyberinfrastructure makes the development and deployment of bioinformatics applications easier by providing the framework and components that may be loosely coupled using service oriented architecture. Here we describe an end to end prototype environment that allows existing applications to run on the grid, taking advantage of open source software that provides a portal interface using GridSphere, with transparent GSI authentication using GAMA, a web service wrapper using Opal, a metascheduler using CSF4, a virtual filesystem using Gfarm, and a grid-enabled cluster environment using Rocks. Solutions to complex problems may be developed using workflow tools such as Kepler that coordinate different interoperable services. The availability of this type of Cyberinfrastructure suggests that new applications should be designed with the grid in mind, using service oriented architecture for interoperability and efficiency. This approach may enable bioinformaticians to focus on their problems of interest, and make use of the emerging Cyberinfrastructure through virtually "drag and drop" deployment. © 2006 IEEE.

DOI： 10.1109/ccgrid.2006.1630932

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/ccgrid/ccgrid2006.html#LiKMIDDSMDWTA06

Language：English   Publishing type：Research paper (international conference proceedings)  

Objectives: In our research on brain function analysis, users require two different simultaneous types of processing: interactive processing to a specific part of data and high-performance batch processing to an entire dataset. The difference between these two types of processing is in whether or not the analysis is for data in the region of interest (ROI). In this study, we propose a Grid portal that has a mechanism to freely assign computing resources to the users on a Grid environment according to the users' two different types of processing requirements. Methods: We constructed o Grid portal which integrates interactive processing and batch processing by the following two mechanisms. First, a job steering mechanism controls job execution based on user-tagged priority among organizations with heterogeneous computing resources. Interactive jobs are processed in preference to batch jobs by this mechanism. Second, a priority-based result delivery mechanism that administrates a rank of data significance. Results: The portal ensures a turn-around time of interactive processing by the priority-based job controlling mechanism, and provides the users with quality of services (QoS) for interactive processing. The users can access the analysis results of interactive jobs in preference to the analysis results of batch jobs. The Grid portal has also achieved high-performance computation of MEG analysis with batch processing on the Grid environment. Conclusion: The priority-based job controlling mechanism has been realized to freely assign computing resources to the users' requirements. Furthermore the achievement of high-performance computation contributes greatly to the overall progress of brain science. The portal has thus made it possible for the users to flexibly include the large computational power in what they want to analyze. © 2005 Schattauer GmbH.

DOI： 10.1055/s-0038-1633960

Scopus

PubMed

researchmap

Language：Japanese   Publisher：Information Processing Society of Japan (IPSJ)  

In these days, biological simulation has been recognized as combinations with simulations at various scales from microscopic to macroscopic, raising our expectation of paradigm shift at computational chemistry. We recognize simulations as programs that calculate on individual time steps repeatedly and then retrieve time continuous results. Thus, a model is proposed for a flow of a time step composed of analysis, monitoring, visualization and data management. We constructed distributed function environment for simulation on OGSA.

CiNii Books

researchmap

researchmap

Language：Japanese   Publishing type：Research paper (other academic)  

researchmap

Language：Japanese   Publisher：Information Processing Society of Japan (IPSJ)  

A scheduling system that can efficiently complete the comoutation of a problem in the class of parallel and cooperative jobs is now demanded among scientists and researchers. However, many solutions for efficiently completing on the Grid mostly target a class of distributed jobs until today. In this paper, a scheduling system whose information-processing mechanism automatically considers dynamic status of computer resources and a flexible exoression based on users' policy are proposed. And its efficiency is demonstrated by using a brain function analysis application which belongs to class of parallel and coooerative jobs.

CiNii Books

researchmap

Language：Japanese   Publisher：Information Processing Society of Japan (IPSJ)  

Scientific data acquired from advanced measurement devices have increased in amount due to the recent development of measurement technologies. To efficiently analyze such data within a realistic time, the Grid is seen as an important computing platform for advancement of science. This paper discusses a Grid-enabled analysis system for brain function. The system enables medical doctors and scientists to analyze a large amount of data acquired from magnetoencephalography, a highly-sophisticated medical device with high efficiency. The feature of this system is in a job-sequence control mechanism based on user's region of interest for data. In this paper, we describe the design and implementation of the system and then evaluate the system qualitatively.

CiNii Books

researchmap

Language：English   Publishing type：Research paper (other academic)  

researchmap

researchmap

researchmap