• Journal of KIISE:Computer Systems and Theory
• /
• v.37 no.4
• /
• pp.246-255
• /
• 2010

There exist virtual machine scheduling delays in virtualized environment in which virtual machines share physical resources. Many resource management systems have been proposed to provide better application QoS through monitoring and analyzing application performance and resource utilization of virtual machines. However, those management systems don't consider virtual machine scheduling delays, result in incorrect application performance evaluation and QoS violations In this paper, we propose an application behavior analysis considering the scheduling delays, and a virtual machine priority allocation scheme based on the analysis to improve the application response time by minimizing the overall virtual machine scheduling delays.

• Proceedings of the Korean Information Science Society Conference
• /
• 2003.10a
• /
• pp.421-423
• /
• 2003

최근 클러스터 웹 서버의 로드 벨런싱으로서 L4 스위치가 널리 사용되고 있다. 전형적인 L4 스위치는 연결 요구 발생시 back-end 노드들의 부하를 평가하여 적은 부하를 갖고 있는 노드로 요청을 보내는 기능을 수행한다. 그러나 L4 스위치의 부하 분산은 사용자 연결 횟수를 이용해 분산함으로서 사용자의 다양한 컨텐츠 요청에 적절한 부하분산 스케줄링을 적용하고 있지 않다. 본 논문에서는 서버에 전송되는 컨텐츠 요구를 L4 스위치 계층에서 인식하여 부하배분에 사용하는 부하분산 모델을 제시한다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.18 no.1
• /
• pp.97-102
• /
• 2014

Abdelzaher et al. proposed an algorithm to determine the schedulability of aperiodic tasks on multiprocessor systems, and proved that the aperiodic tasks are schedulable if the upperbound of synthetic utilization is less than or equal to 0.59. But this algorithm has a drawback in that if some tasks, even though they are completed and have no more execution times, are included in the current invocation set, their execution times and deadlines are added to the synthetic utilization. This may lead to a problem in which actually schedulable tasks are decided not to be schedulable. In this paper, we recognize the above mentioned problem and propose an improved synthetic utilization method that can be used to schedule aperiodic tasks more efficiently on multiprocessor systems.

• Journal of Digital Contents Society
• /
• v.9 no.3
• /
• pp.441-448
• /
• 2008

Recently, several algorithms for scheduling aperiodic tasks have been proposed. Among them, Abdelzaher et al proposed an algorithm to determine the schedulability of aperiodic tasks, and proved that the aperiodic tasks are schedulable if the upperbound of synthetic utilization is less than or equal to $\frac 1{1+\sqrt{1/2}}{\approx}0.59$. But this algorithm has a drawback in that if some tasks, even though they are completed and have no more execution times, are included in the current invocation set, their execution times and deadlines are added to the synthetic utilization. This may lead to a problem in which actually schedulable tasks are decided not to be schedulable. In this paper, we recognize the above mentioned problem and propose an improved synthetic utilization method that can be used to schedule aperiodic tasks more efficiently.

• The Journal of Korean Institute of Next Generation Computing
• /
• v.13 no.3
• /
• pp.72-83
• /
• 2017

GPUs can provide high throughput with latency hiding by executing many warps in parallel. MSHR(Miss Status Holding Registers) for L1 data cache tracks cache miss requests until required data is serviced from lower level memory. In recent GPUs, excessive requests for cache resources cause underutilization problem of GPU resources due to cache resource reservation fails. In this paper, we propose a new warp scheduling technique to reduce stall cycles under MSHR resource shortage. Cache miss rates for each warp is predicted based on the observation that each warp shows similar cache miss rates for long period. The warps showing low miss rates or computation-intensive warps are given high priority to be issued when MSHR is full status. Our proposal improves GPU performance by utilizing cache resource more efficiently based on cache miss rate prediction and monitoring the MSHR entries. According to our experimental results, reservation fail cycles can be reduced by 25.7% and IPC is increased by 6.2% with the proposed scheduling technique compared to loose round robin scheduler.

• Proceedings of the Korean Information Science Society Conference
• /
• 2003.04d
• /
• pp.379-381
• /
• 2003

최근 들어, 인터넷 사용자의 폭발적인 증가로 인하여 차별화된 웹 서비스를 제공해주는 웹 응용프로그램들의 개발이 활발해지고 있다. 이에 따라 웹 서버내의 품질향상을 보장해주는 웹 QoS 기술은 전자상거래나 웹 호스팅 같은 부분에서 점점 더 중요한 문제로 대두되고 있다. 그러나, 대부분의 웹 서버들은 F1F0 방식의 최선 서비스만을 제공하고 있으며, 정보의 중요도나 정보를 제공받는 사용자의 중요도에 따라 차별화된 품질보장을 제공하지 못한다. 본 논문에서는 웹 서비스의 차별화된 품질보장을 제공하는 웹 서버 구현을 위한 두 가지 접근 방식을 제시한다. 첫째는 커널 수준 접근방법으로, 커널상에 실시간 스케줄링 프로세서를 두어 웹 서버에서 수행중인 스케줄링 프로세서와 연동시켜 커널 내부에서도 웹 서버에서 할당된 사용자 요청 우선순위를 유지하도록 한다. 둘째는 부하분산 접근방법으로, IP 수준의 가장법과 터널링 기술을 이용하여 웹 서버의 부하를 분산하여 웹 서비스의 신뢰성을 보장하고 응답속도를 개선한다.

• Journal of KIISE:Computing Practices and Letters
• /
• v.5 no.4
• /
• pp.405-415
• /
• 1999

웹의 성장은 점차 복잡한 응용에 대한 요구를 가중시켰으며, 데이타베이스와 웹과의 효율적인 연동 역시 중요한 문제가 되었다. 본 논문에서는 이와 같은 환경을 효율적으로 지원하기 위한 웹 응용 서버 WATS를 설계하고 구현하였다. WATS에서 응용 프로그램은 컴포넌트 단위로 작성된 후 동적으로 링크되고 요청을 처리할 응용 서버가 대기 상태로 존재하는 확장 API 응용 서버 방식으로 구현되었으며, 이러한 웹 서버와 응용 서버의 분리 구조는 대량의 요청을 처리하기에 적합하다. 또한 컴포넌트를 특성에 따라 분류한 후 각 특성에 적합한 특성 기반 스케줄링 알고리즘을 적용하여 프로세스 부하 균형을 이루었으며, 이 기법이 일반적인 라운드 로빈 스케줄링 알고리즘보다 좋은 성능을 나타냄을 성능 측정을 통해서 보인다.Abstract The increasing popularity of the World-Wide-Web (WWW) has resulted in demand for more complex applications, and web gateways to database became core component in such applications. In this paper, we have designed and implemented WATS in order to support these environments. In WATS, application components are dynamically linked with application server processes. And it is implemented as extensible API application server architecture and is able to process a large amount of requests through separating web server from application server. Also we classify the components into various categories according to its own properties and devise process load balancing algorithm by using property-based scheduling. We show WATS using this algorithm performs better than those using general round-robin algorithm.

• Journal of Advanced Navigation Technology
• /
• v.14 no.6
• /
• pp.865-873
• /
• 2010

In this paper, a novel schedulability criteria is developed to provide handover calls with Quality of Service (QoS) guarantees in terms of both minimum available bandwidth, maximum tolerated packet delay, and other additive QoS constraints as required by the real-time mobile traffic. This requires prediction of the handover time using mobility trends on the mobile station, which is used as input to this work. After the handover time and the QoS are negotiated, the destination base station makes attempts to give priority to handover calls over new calls, and pre-reserves resources that will have more chance of being available during the actual handover.

• Proceedings of the Korean Information Science Society Conference
• /
• 2003.04d
• /
• pp.466-468
• /
• 2003

폭발적인 웹 사용자의 증가를 수용하기 위해서 가격 대 성능비가 우수한 웹 클러스터링 시스템이 선호되고 있다. 많은 연구 결과들은 웹 요청의 패턴이 소수의 사용자 요청에서 매우 높은 확률과 빈도로 참조됨을 보여주고 있다. 이러한 참조 패턴을 반영하기 위해서 서버 시스템은 각각의 사용자 요청의 특성에 맞는 스케줄링 방법을 제공해야 한다. 본 논문에서는 TCP-handoff protocol을 이용한 Layer-7 기반의 클러스터링 시스템을 Linux에 구현하였고, 웹 사용자 요청 패턴에 기반을 둔 DS(Dual Scheduling) 부하 분산 알고리즘을 적용하여 기존의 부하 분산 알고리즘과 비교하였다. 실험적으로 DS 알고리즘은 기존의 알고리즘에 비해 35% 이상의 성능향상을 보여준다.

• Journal of the Korea Society of Computer and Information
• /
• v.16 no.12
• /
• pp.131-137
• /
• 2011

Operating Systems is a discipline which handles complex and abstract concepts related to its components and the principles of how an operating system works. However, most of the OS courses have been textbook-oriented theoretical classes. For students who are familiar with various kinds of media, theoretical classes lead to a decline in the understanding of a lecture and difficulty concentrating. They have tried to make use of educational tools to help students understand a lecture and arouse their interests consistently. This paper describes the design and implementation of a disk scheduling simulator which shows the service processes of disk I/O requests visually. The disk scheduling simulator can be used for demonstrations of the disk scheduling algorithms. The results of the academic achievement evaluation and survey showed that a disk scheduling simulator is useful as an educational tool which causes the interests about operating systems and enhances the understanding of a lecture.