• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.12 no.3
• /
• pp.385-396
• /
• 1999

In the process of structural design for large-scale structures with several thousands of degrees of freedom, a plethora of structural calculations with large amount of data storage are required to obtain the forces and displacements of the members. However, current computational environment with single microprocessor such as a personal computer or a workstation is not capable of generating a high-level of efficiency in structural analysis and design process for large-scale structures. In this paper, a high-performance parallel computing system interconnected by a network of personal computers is proposed for an efficient structural analysis. Two distributed structural analysis algorithms are developed in the form of distributed or parallel preconditioned conjugate gradient (DPCG) method. To enhance the performance of the developed distributed structural analysis algorithms, the number of communications and the size of data to be communicated are minimized. These algorithms are applied to the structural analyses of three large space structures as well as a 144-story tube-in-tube framed structure.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.6 no.11
• /
• pp.2827-2848
• /
• 2012

Previously, we described a social media cloud computing service environment (SMCCSE). This SMCCSE supports the development of social networking services (SNSs) that include audio, image, and video formats. A social media cloud computing PaaS platform, a core component in a SMCCSE, processes large amounts of social media in a parallel and distributed manner for supporting a reliable SNS. Here, we propose a Hadoop-based multimedia system for image and video transcoding processing, necessary functions of our PaaS platform. Our system consists of two modules, including an image transcoding module and a video transcoding module. We also design and implement the system by using a MapReduce framework running on a Hadoop Distributed File System (HDFS) and the media processing libraries Xuggler and JAI. In this way, our system exponentially reduces the encoding time for transcoding large amounts of image and video files into specific formats depending on user-requested options (such as resolution, bit rate, and frame rate). In order to evaluate system performance, we measure the total image and video transcoding time for image and video data sets, respectively, under various experimental conditions. In addition, we compare the video transcoding performance of our cloud-based approach with that of the traditional frame-level parallel processing-based approach. Based on experiments performed on a 28-node cluster, the proposed Hadoop-based multimedia transcoding system delivers excellent speed and quality.

• KIISE Transactions on Computing Practices
• /
• v.21 no.8
• /
• pp.561-566
• /
• 2015

Hadoop is becoming widely adopted in scientific and commercial areas as an open-source distributed data processing framework. Recently, for real-time processing and analysis of data, an attempt to apply high-performance computing technologies to Hadoop is being made. In this paper, we have expanded the Hadoop Filesystem library to support Lustre, which is a popular high-performance parallel distributed filesystem, and implemented the Hadoop MapReduce execution environment over the Lustre filesystem. We analysed Hadoop MapReduce over Lustre by using Hadoop standard benchmark tools. We found that Hadoop MapReduce over Lustre execution has a performance 2-13 times better than a typical Hadoop MapReduce execution.

• The KIPS Transactions:PartA
• /
• v.14A no.2
• /
• pp.99-106
• /
• 2007

Grid computing, a mechanism which uses heterogeneous systems that are geographically distributed, draws attention as a new paradigm for the next generation operation of parallel and distributed computing. The importance of grid computing concerning communication cost is very huge because grid computing furnishes uses with integrated virtual computing service, in which a number of computer systems are connected by a high-speed network. Therefore, to reduce the execution time, the scheduling algorithm in grid environment should take communication cost into consideration as well as computing ability of resources. However, most scheduling algorithms have not only ignored the communication cost by assuming that all tasks were dealt in one cluster, but also did not consider the overhead of communication cost when the tasks were processed in a number of clusters. In this paper, the functions of original scheduling algorithms are analyzed. More importantly, the functions of algorithms are compared and analyzed with consideration of communication cost within the co allocation environment, in which a task is performed separately in many clusters.

• Journal of KIISE:Computer Systems and Theory
• /
• v.27 no.6
• /
• pp.608-618
• /
• 2000

We consider the problem of placing resources in a distributed computing system so that certain performance requirements may be met while minimizing the number of required resource copies, irrespective of node or link failures. To meet the requirements for high performance and high availability, minimum number of resource copies should be placed in such a way that each node has at least two copies on the node or its neighbor nodes. This is called the fault-tolerant resource placement problem in this paper. The structure of a parallel or a distributed computing system is represented by a graph. The fault-tolerant placement problem is first transformed into the problem of finding the smallest fault-tolerant dominating set in a graph. The dominating set problem is known to be NP-complete. In this paper, searching for the smallest fault-tolerant dominating set is formulated as a state-space search problem, which is then solved optimally with the well-known A* algorithm. To speed up the search, we derive heuristic information by analyzing the properties of fault-tolerant dominating sets. Some experimental results on various regular and random graphs show that the search time can be reduced dramatically using the heuristic information.

• Journal of Information Technology Services
• /
• v.8 no.1
• /
• pp.193-202
• /
• 2009

Cloud computing emerges as a new computing paradigm which targets reliable and customizable services. The term builds on decades of research in virtual machine, distributed and parallel computing, utility computing, and more recently networking, web service, and software as a service. In this paper, we provide a seamless connection migration of web services. This is useful for cloud computing environment in which many client terminals have mobility. With the wireless internet facility, those mobile users can move place to place during internet communication. Therefore, we provide solutions to the two major problems in current virtualization based migration: communication failure problems and connection re-establishment. Communication channel flushing by zero window notification helps to resolve the communication failure problems and TCP port inheritance prevents connection re-establishment errors during socket reconstruction. Thus, our web service migration facility is now able to preserve open network connections, and even for server sockets. This is a highly transparent approach, in that we did not Introduce additional messages for channel flushing and did not make any modification to the TCP protocol stack. Experimental results show that the overhead due to connection migration of web services is almost negligible when compared with time to take the conventional web service migration.

• Journal of Internet Computing and Services
• /
• v.18 no.2
• /
• pp.53-60
• /
• 2017

Since processors are handling inputs and outputs independently on distributed memory computers, different file input/output methods are used. In this paper, we implemented and compared various file I/O methods to show their efficiency on distributed memory parallel computers. The implemented I/O systems are as following: (i) parallel I/O using NFS, (ii) sequential I/O on the host processor and domain decomposition, (iii) MPI-IO. For performance analysis, we used a separated file server and multiple processors on one or two computational servers. The results show the file I/O with NFS for inputs and sequential output with domain composition for outputs are best efficient respectively. The MPI-IO result shows unexpectedly the lowest performance.

• The Transactions of the Korea Information Processing Society
• /
• v.6 no.7
• /
• pp.1777-1788
• /
• 1999

The portability of Java language makes it possible to use heterogeneous computers without re-compiling of application programs. Java applet can also be transported to other computers via Web browser. In this research, a Cooperative Web Computing Environment(CWCE) that uses idle computers on the Intranet for cooperative parallel computing work is suggested. The CWCE allows to use more than a manager computer that sends applets and manages communication between other computers. The number of manager computers can be determined according to the characteristics of computing environment and any chosen application program. It can reduce the amount of communication overhead for the application programs especially with synchronized communication. For the CWCE, a decision function to determine the managing level is provided. The CWCE turns out to be useful computing environment for the applications with less computation request ratio and multi-managing can help to reduce the communication overhead especially for the applications with a high ratio of synchronization purpose communications.

• Proceedings of the Computational Structural Engineering Institute Conference
• /
• 2000.10a
• /
• pp.40-47
• /
• 2000

One of the drawbacks of GA-based structural optimization is that the fitness evaluation of a population of hundreds of individuals requiring hundreds of structural analyses at each CA generation is computational too expensive. Therefore, a parallel genetic algorithm is developed for structural optimization on a cluster of personal computers in this paper. Based on the parallel genetic algorithm, a population at every generation is partitioned into a number of sub-populations equal to the number of slave computers. Parallelism is exploited at sub-population level by allocationg each sub-population to a slave computer. Thus, fitness of a population at each generation can be concurrently evaluated on a cluster of personal computers. For implementation of the algorithm a virtual distributed computing system in a collection of personal computers connected via a 100 Mb/s Ethernet LAN. The algorithm is applied to the minimum weight design of a steel structure. The results show that the computational time requied for serial GA-based structural optimization process is drastically reduced.

• IEIE Transactions on Smart Processing and Computing
• /
• v.5 no.4
• /
• pp.289-293
• /
• 2016

As more mobile devices are equipped with multi-core CPUs and are required to execute many compute-intensive multimedia applications, it is important to optimize the systems, considering the underlying parallel hardware architecture. In this paper, we implement and optimize ray-tracing application tailored to a given mobile computing platform with multiple heterogeneous processing elements. In this paper, a lightweight ray-tracing application is specified and implemented in Kahn process network (KPN) model-of-computation, which is known to be suitable for the description of real-time applications. We take an open-source C/C++ implementation of ray-tracing and adapt it to KPN description in the Distributed Application Layer framework. Then, several possible configurations are evaluated in the target mobile computing platform (Exynos 5422), where eight heterogeneous ARM cores are integrated. We derive the optimal degree of parallelism and a suitable distribution of the replicated tasks tailored to the target architecture.