• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2004년도 학술대회지
• /
• pp.608-611
• /
• 2004

Grid computing environment can be used to interconnect a wide variety of geographically distributed heterogeneous computing resources based on high-speed network. To make business service, it is necessary for Grid accounting system to measure the computational cost by consuming computer resources. To collect resource consumption data, and to keep track of process without needing to recompile kernel source, we use system call wrapping. By making use of this technique, we modifies system call table and replace existing system call to new system call that can monitor processes running in CPU kernel currently. Therefore we can measure user process resource usage for Grid accounting system.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제10권12호
• /
• pp.5174-5190
• /
• 2016

Recent years have witnessed an explosive growth of mobile devices, mobile cloud computing services offered by these devices and the remote clouds behind them. In this paper, we noticed ultra-low latency service, as a type of mobile cloud computing service, requires extremely short delay constraints. Hence, such delay-sensitive applications should be satisfied with strong QoS guarantee. Existing solutions regarding this problem have poor performance in terms of throughput. In this paper, we propose an end-to-end bandwidth resource reservation via software defined scheduling inspired by the famous SDN framework. The main contribution of this paper is the end-to-end resource reservation and flow scheduling algorithm, which always gives priority to delay sensitive flows. Simulation results confirm the advantage of the proposed solution, which improves the average throughput of ultra-low latency flows.

• Journal of Information Processing Systems
• /
• 제18권3호
• /
• pp.389-401
• /
• 2022

The business of Internet of Vehicles (IoV) is growing rapidly, and the large amount of data exchange has caused problems of large mobile network communication delay and large energy loss. A strategy for resource allocation of IoV communication based on mobile edge computing (MEC) is thus proposed. First, a model of the cloud-side collaborative cache and resource allocation system for the IoV is designed. Vehicles can offload tasks to MEC servers or neighboring vehicles for communication. Then, the communication model and the calculation model of IoV system are comprehensively analyzed. The optimization objective of minimizing delay and energy consumption is constructed. Finally, the on-board computing task is coded, and the optimization problem is transformed into a knapsack problem. The optimal resource allocation strategy is obtained through genetic algorithm. The simulation results based on the MATLAB platform show that: The proposed strategy offloads tasks to the MEC server or neighboring vehicles, making full use of system resources. In different situations, the energy consumption does not exceed 300 J and 180 J, with an average delay of 210 ms, effectively reducing system overhead and improving response speed.

• Journal of Information Processing Systems
• /
• 제20권2호
• /
• pp.226-238
• /
• 2024

Recently, multi-access edge computing (MEC) has emerged as a promising technology to alleviate the computing burden of vehicular terminals and efficiently facilitate vehicular applications. The vehicle can improve the quality of experience of applications by offloading their tasks to MEC servers. However, channel conditions are time-varying due to channel interference among vehicles, and path loss is time-varying due to the mobility of vehicles. The task arrival of vehicles is also stochastic. Therefore, it is difficult to determine an optimal offloading with resource allocation decision in the dynamic MEC system because offloading is affected by wireless data transmission. In this paper, we study computation offloading with resource allocation in the dynamic MEC system. The objective is to minimize power consumption and maximize throughput while meeting the delay constraints of tasks. Therefore, it allocates resources for local execution and transmission power for offloading. We define the problem as a Markov decision process, and propose an offloading method using deep reinforcement learning named deep deterministic policy gradient. Simulation shows that, compared with existing methods, the proposed method outperforms in terms of throughput and satisfaction of delay constraints.

• 한국콘텐츠학회:학술대회논문집
• /
• 한국콘텐츠학회 2007년도 추계 종합학술대회 논문집
• /
• pp.226-230
• /
• 2007

우리나라의 컴퓨터 보유율과 인터넷사용률은 세계1위를 자랑한다. 2002년 이래로 운영되어 오고 있는 코리아앳홈(이하 Korea@Home)프로젝트는 우리나라가 보유한 풍부한 컴퓨팅 자원들을 모아서 대규모 응용에 활용하고자 하는 것이다. 이미 많은 회원들을 확보하고 있으며, 현재 신약개발과 기후예측 등의 응용을 수행하고 있다. 코리아앳홈 회원들이 자신의 컴퓨터 자원을 제공하는데 있어 사용자들에게 불편을 주지 않기 위해 자원제공자들의 기여 패턴을 분석하여 사용자가 컴퓨터를 사용하지 않는 시간에 따라 에이전트t의 작업을 수행하도록 하는 스케줄링 방법을 제안한다

• 한국컴퓨터정보학회논문지
• /
• 제11권4호
• /
• pp.69-75
• /
• 2006

과학 기술이 발전함에 따라 대량의 정보를 처리하기 위해 대두된 인터넷을 기반으로 하는 분산 컴퓨팅 환경은 대규모의 독립된 자원을 공유하여 과학 연구와 같은 문제를 해결하기 위한 구축된 환경이므로, 사용자 작업을 효율적으로 할당하기 위한 스케줄링 알고리즘이 필요하다. 현재까지 여러 스케줄링 알고리즘이 연구되어 왔지만, 대부분 피어의 자율성을 고려하지 않는 문제점을 가지고 있다. 본 논문에서는 이러한 문제점을 해결하기 위하여 인터넷기반 분산 컴퓨팅 환경에서의 피어 가용길이 예상 기법을 제안하였다. 또한 인터넷기반 분산 컴퓨팅환경에서 사용되는 SRTFIT 알고리즘에 적용하여, 시뮬레이션을 통하여 제안된 기법이 단순한 예상기법보다 성능이 뛰어남을 보였다.

• 정보처리학회논문지:컴퓨터 및 통신 시스템
• /
• 제6권12호
• /
• pp.479-486
• /
• 2017

볼런티어 컴퓨팅(Volunteer Computing)은 많은 노드들의 유휴자원을 이용하여 연산을 수행하는 새로운 컴퓨팅 패러다임이다. 볼런티어 컴퓨팅 수행을 위해 운영하는 클라이언트 어플리케이션은 사용자의 설정 정보에 의해 동작 방식이 결정된다. 이상적인 동작을 위해서는 시스템 특징과 다른 어플리케이션의 동작 방식에 최적화된 설정이 요구된다. 본 연구에서는 유휴 자원 정보를 주기적으로 CPU와 GPU의 사용 비율을 분석하고 최적화된 옵션을 정해 동적으로 적용하는 관리자를 개발하였다. 또한 CPU 자원의 높은 활용도를 위해 태스크 스케일링을 진행하고 CPU코어를 주기적으로 재 할당 하여 CPU자원이 균등하게 사용되게 하였다. 제시하는 기법을 통해 기존의 볼런티어 컴퓨팅보다 높은 태스크 연산 능력을 보였으며 성능간섭 또한 최소화 시켰다. 볼런티어 컴퓨팅을 진행하는데 있어 볼런티어들이 더 높은 컴퓨팅 자원을 제공할 수 있게 될 것으로 예상한다.

• 한국정보과학회논문지:정보통신
• /
• 제31권1호
• /
• pp.112-122
• /
• 2004

본 논문에서는 이동 컴퓨팅 환경에서 셀 내부 영역에 기반을 둔 이동성 예측 모델과 모바일 호스트들의 이동성을 토대로 하는 이동성 등급을 동적으로 조정할 수 있는 자원 예약 기법을 제안한다. 본 논문에서 제안하는 이동성 예측 모델은 셀 내부의 각각 다른 영역을 기반으로 가약 마코프 체인을 이용하여 모델링 하였으며, 새로운 방법으로 핸드오프 확률을 추정한다. 그리고 모바일 호스트들이 셀 내부에 머물 가능성이 있는 셀 내부 존재 시간을 제공하며, 그 셀 내부 존재 시간은 동적인 자원 예약 기법의 새로운 인자로 활용된다. 동적인 자원 예약 기법은 모바일 호스트들의 내부 존재 시간의 마감에 따라서 변화하는 이동성 등급을 이용하여 이웃 셀의 자원을 동적으로 예약한다. 이 기법들은 신규 호 블록킹율과 핸드 오프 호 종료율을 개선시켰다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제5권1호
• /
• pp.83-104
• /
• 2011

Mobile grid is a combination of grid computing and mobile computing to build grid systems in a wireless mobile environment. The development of network technology is assisting in realizing mobile grid. Mobile grid based on established grid infrastructures needs effective resource management and reliable job scheduling because mobile grid utilizes not only static grid resources but also dynamic grid resources with mobility. However, mobile devices are considered as unavailable resources in traditional grids. Mobile resources should be integrated into existing grid sites. Therefore, this paper presents a mobile grid middleware interconnecting existing grid infrastructures with mobile resources and a mobile service agent installed on the mobile resources. This paper also proposes a mobile resource reliability-based job scheduling model in order to overcome the unreliability of wireless mobile devices and guarantee stable and reliable job processing. In the proposed job scheduling model, the mobile service agent calculates the mobile resource reliability of each resource by using diverse reliability metrics and predicts it. The mobile grid middleware allocated jobs to mobile resources by predicted mobile resource reliability. We implemented a simulation model that simplifies various functions of the proposed job scheduling model by using the DEVS (Discrete Event System Specification) which is the formalism for modeling and analyzing a general system. We also conducted diverse experiments for performance evaluation. Experimental results demonstrate that the proposed model can assist in improving the performance of mobile grid in comparison with existing job scheduling models.

• 한국BIM학회 논문집
• /
• 제10권3호
• /
• pp.43-53
• /
• 2020

BIM-based Cloud platform gained popularity coupled with the convergence of Fourth Industrial Revolution technology. However, most of the previous work has not guaranteed sufficient efficiency to meet user requirements according to BIM service. Furthermore, the Cloud environment is only used as a server and it does not consider cloud characteristics. For the processing of High Capacity Data like BIM and using seamless BIM service, Resource management technology is required in the cloud environment. In this paper, to solve the problems, we propose a BIM platform for BaaS and an efficient resource allocation scheme. We also proved the efficiency of resource for the proposed scheme by using existing schemes. By doing this, the proposed scheme looks forward to accelerating the growth of the BaaS through improving the user experience and resource efficiency.