• Journal of Internet Computing and Services
• /
• v.18 no.6
• /
• pp.47-53
• /
• 2017

Recently, huge datasets have been generating rapidly in a variety of fields. Then, there is an urgent need for technologies that will allow efficient and effective processing of huge datasets. Therefore the problems of partitioning a huge dataset effectively and alleviating the processing overhead of the partitioned data efficiently have been a critical factor for scalability and performance in distributed database system. In our work we utilized multicore servers to provide scalable service to our distributed system. The partitioning of database over multicore servers have emerged from a need for new architectural design of distributed database system from scalability and performance concerns in today's data deluge. The system allows uniform access through a web service interface to concurrently distributed databases over multicore servers, using SQMD (Single Query Multiple Database) mechanism based on publish/subscribe paradigm. We will present performance results with the distributed database system built on multicore server, which is time intensive with traditional architectures. We will also discuss future works.

• Journal of the Korea Society for Simulation
• /
• v.8 no.2
• /
• pp.1-11
• /
• 1999

Advanced information communication processing system mainly consists of network access subsystems and a switching system. This paper provides performance analysis of a typical network access subsystem. The network access subsystem is modeled as a queueing network including a server providing polling services. The arrival process of messages to an input buffer is regarded as a Poisson process. Performance measures such as mean input buffer length and mean waiting time of meassages are obtained through simulation, for it is impossible to calculate the performance measures using an analytical method. Hybrid simulation is used to reduce the variance of estimators. The variance reduction effect on the mean waiting time is reported.

• Journal of information and communication convergence engineering
• /
• v.19 no.4
• /
• pp.257-262
• /
• 2021

Providing web services to users has become expensive in recent times. For better web services, a web server is provided with high-performance technology. To achieve great web service experiences, tools such as general-purpose graphics processing units (GPGPUs), artificial intelligence, high-performance computing, and three-dimensional simulation are widely used. However, graphics processing units (GPUs) are used in high-speed operations and have limited general applications. In this study, we developed a task queue in a GPU to improve the performance of a web service using a multiprocessor and studied how to receive and process user requests in bulk. We propose the use of a GPGPU-based task queue to process user requests more than GPGPU based a central processing unit thread, and to process more GPU threads on task queue at about 136% to 233%, and proved that the proposed method is effective for web service.

• Journal of KIISE:Computing Practices and Letters
• /
• v.16 no.11
• /
• pp.1041-1045
• /
• 2010

Clusters can provide scalable and flexible architectures for parallel computing servers and data centers. Their performance prediction has been a very challenging issue. Existing performance measurement methodologies are able to measure the performance of servers already constructed. Thus they cannot provide a way to predict the overall system performance in advance when designing the system at the initial phase or adding more nodes for more capacity. Therefore, the performance modeling and prediction methodology for large-scale clusters is highly required. In this paper, we suggest a methodology to predict the performance of large-scale clusters, which consists of measurement, modeling and prediction steps. We apply the methodology to a real cluster server and show its usefulness.

• KIISE Transactions on Computing Practices
• /
• v.24 no.1
• /
• pp.10-16
• /
• 2018

The task of consolidating server workloads is critical for the efficiency of a datacenter in terms of reducing costs. However, as a greater number of workloads are consolidated in a single server, the performance of workloads might be degraded due to their contention to the limited shared resources. To reduce the performance degradation, scheduling for mitigating the contention of shared resources is necessary. In this paper, we present the Dynamic Voltage Frequency Scaling (DVFS) based memory-contention aware scheduling method for multi-threaded workloads. The proposed method uses two approaches: running memory-intensive threads on the limited cores to avoid concurrent memory accesses, and reducing the frequencies of the cores that run memory-intensive threads. With the proposed algorithm, we increased performance by 43% and reduced power consumption by 38% compared to the Completely Fair Scheduler(CFS), the default scheduler of Linux.

• Journal of KIISE:Computer Systems and Theory
• /
• v.26 no.11
• /
• pp.1296-1304
• /
• 1999

본 논문은 다수의 고객들에게 실시간 재생 서비스를 제공하는 주문형 비디오 서버에서 데이타 검색의 효율을 높이기 위한 저장 시스템의 구조와 그에 따른 배치 정책을 제안한다. 주문형 비디오 저장 서버에서 동시에 보다 많은 고객들에게 그들이 원하는 서비스를 제공하기 위해서는 여러 분야에서의 다양한 정책들이 고려될 수 있으나, 특히 저장 매체들을 어떤 구조로 구성하고 여기에 비디오 객체들을 어떻게 배치할 것인가는 검색 효율과 직접 관계되는 중요한 문제이다. 본 논문에서는 디스크 배열 형태로 구성된 저장 시스템을 디스크 성능을 고려하여 재구성하고, 비디오 객체들을 저장할 때 저장하고자 하는 데이타의 특성들과 저장 구조를 함께 고려하여 검색 효율을 극대화할 수 있는 배치 방법을 제안한다. 마지막으로 제안된 정책의 검색 효율을 검증하기 위해 다양한 실험을 통하여 기존의 배치 정책들과 비교하고 성능을 평가한다.Abstract In this paper, we propose the structure of storage system and a placement policy to provide many clients with real-time playback services efficiently in VOD(video-on-demand) server. Though policies in various areas being considered to provide more clients with services of degree requested by them simultaneously in VOD storage server, it is important how to construct storage media and to place video objects on it for retrieval efficiency. In this paper, we reorganize a large disk array with disks performance and place video objects using the placement policy considering both characteristics of video data and the structure of storage system for maximizing retrieval efficiency. Lastly, we simulate the proposed policy and conventional policies through various environments, compare our policy with others n evaluate the performance of our policy.

• Journal of the Korea Computer Industry Society
• /
• v.3 no.5
• /
• pp.529-546
• /
• 2002

A VOD server, which is the central component for constructing VOD systems, requires to provide high bandwidth and continuous real-time delivery. It is also necessary to the sophisticated disk scheduling and data placement schemes in VOD sewers. One of the most common problem facing in such a system is the high initial latency time to service multiple users concurrently. In this paper, we propose a dynamic stream merging scheme for reducing the initial latency time in VOD servers. The proposed scheme allows clients to merge streams on a request as long as their requests fall within the reasonable time interval. The basic idea behind the dynamic stream merging is to merge multiple streams into one by increasing the frame rate of each stream. In the performance study, the proposed scheme can reduce the initial latency time under the minimum buffer use and also can enhance the performance of the VOD server with respect to the capacity of user admission.

• The Transactions of the Korea Information Processing Society
• /
• v.4 no.12
• /
• pp.2971-2984
• /
• 1997

In this paper, we propose an efficient video block Placement scheme that utilize the current disk product that has MZR disk characteristic and users' skewed access pattern on VOD. Also, we evaluate its performance through simulation and modeling of VOD server. The basic placement rule is to place on MZR disks by LP, SHP methods according to the Zipf distribution of popularity. To verify the proposed scheme, we examined its performance on workstation with 2 MZR disks under varied skewed factors. The proposed placement scheme showed better response time than the random method. To extend proposed placement scheme to disk group, we analyzed the theoretical maximum numbers of concurrent users and the required buffer size per user. For performance parameters for the proposed scheme, we considered the disk head scheduling methods, the placement methods and the striping unit. The result of experiments showed that the proposed scheme was effective.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.51 no.9
• /
• pp.103-108
• /
• 2014

In the virtualized server systems, a scheduler in a hypervisor is responsible to assign physical resources for virtual machines. However, the traditional scheduler is hard to provide optimized resource allocation considering the amount of I/O requests. Especially, the drawback hinders performance of software load balancer which runs on virtual machines to distribute I/O requests from the clients. In this paper, we propose a new architecture to improve the performance of software load balancer using multiprocess. Our architecture aims to improve hardware resource utilization and overall performance of the server systems which utilize virtualization technology. Experimental results show the effectiveness of the proposed architecture for the various cases.

• The Transactions of the Korea Information Processing Society
• /
• v.7 no.8
• /
• pp.2310-2326
• /
• 2000

Client-server DBMS based on a data-shipping model can exploit e1ient resources effectively by allowing inter-transaction caching. However, inter-transaction caching raises the need of transactional cache consistency maintenancetTCCM protocol. since each client is able to cache a portion of the database dynamically. Deferred locking(DL) is a new detection-based TCCM scheme designed on the basis of a primary copy locking algorithm. In DL, a number of lock ,ujuests and a data shipping request are combined into a single message packet to minimize the communication overhead required for consistency checking. Lsing a simulation model. the performance of the prolxlsed scheme is compared with those of two representative detection based schemes, the adaptive optimistic concurrency control and the caching two-phase locking. The performance results indicate that DL improves the overall system throughput with a reasonable transaction abort ratio over other detection - based schemes.