• Spatial Information Research
• /
• 제19권4호
• /
• pp.55-71
• /
• 2011

최근 IT 분야에서 인터넷을 기반으로 IT 자원들을 서비스 형태로 제공하는 클라우드 컴퓨팅에 대한 관심이 증대되고 있으며, 이에 따라 대규모 데이터를 수많은 서버들에 분산 저장하고 관리하기 위한 분산 데이터 처리 기법에 대한 연구가 활발히 진행되고 있다. 한편 GIS 기술의 성장과 더불어 급격히 증가하고 있는 공간 데이터를 효율적으로 활용하기 위해서는, 클라우드 컴퓨팅을 이용한 대용량 공간데이터의 분산 처리가 필수적이다. 이를 위해 본 논문에서는 대표적인 분산 데이터 처리 기법에 대해 살펴보고, 분산 데이터 처리 기법 성능 개선을 위한 최적화 요구사항을 분석한다. 마지막으로 Hadoop 기반 클러스터를 구축하고 이를 통해서 분산 데이터 처리 기법의 성능 최적화에 대한 성능평가를 수행한다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제10권9호
• /
• pp.4165-4187
• /
• 2016

As an emerging technology, cloud computing is a revolution in information technology that attracts significant attention from both public and private sectors. In this paper, we proposed a dynamic approach for live migration to obviate overloaded machines. This approach is applied on OpenStack, which rapidly grows in an open source cloud computing platform. We conducted a cost-aware dynamic live migration for virtual machines (VMs) at an appropriate time to obviate the violation of service level agreement (SLA) before it happens. We conducted a preemptive migration to offload physical machine (PM) before the overload situation depending on the predictive method. We have carried out a distributed model, a predictive method, and a dynamic threshold policy, which are efficient for the scalable environment as cloud computing. Experimental results have indicated that our model succeeded in avoiding the overload at a suitable time. The simulation results from our solution remarked the very efficient reduction of VM migrations and SLA violation, which could help cloud providers to deliver a good quality of service (QoS).

• 한국멀티미디어학회논문지
• /
• 제19권5호
• /
• pp.808-817
• /
• 2016

Recently, the wearable computing technology with bio-sensors has been rapidly developed and utilized in various areas such as personal health, care-giving for senior citizens who live alone, and sports activities. In particular, the wearable computing equipment to measure vital signs by means of digital yarns and bio sensors is noticeable. The wearable computing devices help users monitor and manage their health in their daily lives through the customized healthcare service. In this paper, we suggest a system for monitoring and analyzing vital signs utilizing smart healthcare clothing with bio-sensors. Vital signs that can be continuously acquired from the clothing is well-known as unstructured data. The amount of data is huge, and they are perceived as the big data. Vital sings are stored by Hadoop Distributed File System(HDFS), and one can build data warehouse for analyzing them in HDFS. We provide health monitoring system based on vital sings that are acquired by biosensors in smart healthcare clothing. We implemented a big data platform which provides health monitoring service to visualize and monitor clinical information and physical activities performed by the users.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제9권7호
• /
• pp.2649-2666
• /
• 2015

In the dynamic cloud computing environment, to ensure, under the terms of service-level agreements, the maximum efficiency of resource utilization, it is necessary to investigate the online dynamic management of virtual machine resources and their operational application systems/components. In this study, the feasibility and properties of the division of virtual machine resources on the cloud platform, using the virtual machine cluster as the management unit, are investigated. First, the definitions of virtual machine clusters are compared, and our own definitions are presented. Then, the feasibility of division using the virtual machine cluster as the management unit is described, and the isomorphism and reconfigurability of the clusters are proven. Lastly, from the perspectives of clustering and cluster segmentation, the dynamics of virtual machines are described and experimentally compared. This study aims to provide novel methods and approaches to the optimization management of virtual machine resources and the optimization configuration of the parameters of virtual machine resources and their application systems/components in large-scale cloud computing environments.

• 대한임베디드공학회논문지
• /
• 제11권6호
• /
• pp.325-334
• /
• 2016

Recently, Open Computing Language (OpenCL) has been proposed to provide a framework that supports heterogeneous computing platforms. By using an OpenCL framework, digital communication systems can support various protocols in a unified computing environment to achieve both high portability and high performance. This article introduces a parallel software decoder of Low Density Parity Check (LDPC) codes for China Multimedia Mobile Broadcasting (CMMB) on a heterogeneous platform. Each step of LDPC decoding has different parallelization characteristics. In this paper, steps suitable for task-level parallelization are executed on the CPU, and steps suitable for data-level parallelization are processed by the GPU. To improve the performance of the proposed OpenCL kernels for LDPC decoding operations, explicit thread scheduling, loop-unrolling, and effective data transfer techniques are applied. The proposed LDPC decoder achieves high performance by using heterogeneous multi-core processors on a unified computing framework.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제8권5호
• /
• pp.1554-1566
• /
• 2014

Cloud Computing allows application providers seamlessly scaling their services and enables users scaling their usage according to their needs. In this paper, using queuing game model, we present service scheduling schemes which are used in software as a service (SaaS). The object is maximizing the Cloud Computing platform's (CCP's) payoff via controlling the service requests whether to join or balk, and controlling the value of CCP's admission fee. Firstly, we treat the CCP as one virtual machine (VM) and analyze the optimal queue length with a fixed admission fee distribution. If the position number of a new service request is bigger than the optimal queue length, it balks. Otherwise, it joins in. Under this scheme, the CCP's payoff can be maximized. Secondly, we extend this achievement to the multiple VMs situation. A big difference between single VM and multiple VMs is that the latter one needs to decide which VM the service requests turn to for service. We use a corresponding algorithm solve it. Simulation results demonstrate the good performance of our schemes.

• Journal of information and communication convergence engineering
• /
• 제8권4호
• /
• pp.427-432
• /
• 2010

The fast-emerging of cloud computing technology today has sufficiently benefited its wide range of users from individuals to large organizations. It carries an attractive characteristic by renting myriad virtual storages, computing resources and platform for users to manipulate their data or utilize the processing resources conveniently over Internet without the need to know the exact underlying infrastructure which is resided remotely at cloud servers. However due to the loss of direct control over the systems/applications, users are concerned about the risks of cloud services if it is truly secured. In the literature, there are cases where attackers masquerade as cloud users, illegally access to their accounts, by stealing the static login password or breaking the poor authentication gate. In this paper, we propose a two-factor authentication framework to enforce cloud services' authentication process, which are Public Key Infrastructure (PKI) authentication and mobile out-of-band (OOB) authentication. We discuss the framework's security analysis in later session and conclude that it is robust to phishing and replay attacks, prohibiting fraud users from accessing to the cloud services.

• Journal of Electrical Engineering and Technology
• /
• 제13권1호
• /
• pp.181-191
• /
• 2018

Rapid development of information technology makes our environment become smarter and massive high performance computers are providing powerful computing for that. Graphics Processing Unit (GPU) as a typical high performance component is being widely used for both graphics and general-purpose applications. Although it can greatly improve computing power, it also delivers significant power consumption and need sufficient power supplies. To make high performance computing more sustainable, the important step is to measure it. Current power technologies for GPU have some drawbacks, such as they are not applicable for power estimation at the early stage. In this article, we present a novel power technology to correlate power consumption and the characteristics at the programmer perspective, and then to estimate power consumption of source program without prerunning. We conduct experiments on Nvidia's GT740 platform; the results show that our power model is more accurately than regression model and has an average error of 2.34% and the maximum error of 9.65%.

• 한국정보통신학회논문지
• /
• 제14권5호
• /
• pp.1146-1154
• /
• 2010

오픈 소스 클라우드 컴퓨팅 플랫폼인 OpenNebula를 이용한 클라우드 컴퓨팅 환경에서 가상머신 임대 서비스를 구현하고 클라우드 자원 관리와 서비스 사용의 편의성을 위하여 웹기반 클라우드 사용자 인터페이스를 구현하였다. OpenNebula의 가상머신 할당 기법은 가상화 소프트웨어의 CPU 할당 스케줄러를 고려하지 않아 성능 저하의 요인이 되고 있다. 이러한 문제점을 해결하기 위하여 클러스터 노드의 유휴 CPU 자원의 우선순위와 Xen의 Credit 스케줄러를 고려하여 OpenNebula의 가상머신 할당 스케줄러의 성능을 개선하였다. 실험 결과 제안한 가상머신 할당기법은 기존 방식에 비하여 수용 가능한 가상머신 수와 CPU 자원 할당량에서 향상된 결과를 보였다.

• 전기학회논문지
• /
• 제66권1호
• /
• pp.100-106
• /
• 2017

In this paper, we propose a parallel optimization method of Aho-Corasick (AC) algorithm and Parallel Failureless Aho-Corasick (PFAC) algorithm using Open Computing Language (OpenCL) on Field Programmable Gate Array (FPGA). The low throughput of string matching engine causes the performance degradation of network process. Recently, many researchers have studied the string matching engine using parallel computing. FPGA's vendors offer a parallel computing platform using OpenCL. In this paper, we apply the AC and PFAC algorithm on DE1-SoC board with Cyclone V FPGA, where the optimization that considers FPGA architecture is performed. Experiments are performed considering global id, local id, local memory, and loop unrolling optimizations using PFAC algorithm. The performance improvement using loop unrolling is 129 times greater than AC algorithm that not adopt loop unrolling. The performance improvements using loop unrolling are 1.1, 0.2, and 1.5 times greater than those using global id, local id, and local memory optimizations mentioned above.