• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.5
• /
• pp.1606-1623
• /
• 2015

Optical switching will be applied in datacenter networks because electronic switching is costly and power-consuming. In this paper, considering the ring-based interconnection using optical switching in the core of a datacenter, we study the home circuit (HC) construction for the labeled optical burst switching with home circuit (LOBS-HC), a new paradigm trying to share wavelengths among the HCs from the same source. In particular, aiming to construct HCs dynamically and properly, a scheme named optimal path matching and symmetric HC matching (OPM-SHM) is proposed. The main idea of OPM-SHM is to dynamically construct HCs by sharing wavelength(s) not only among the same-source HCs but also with symmetric HCs which have different sources other than the original LOBS-HC features. The simulation results demonstrate that OPM-SHM achieves better performance than some other methods in terms of burst loss rate and wavelength utilization of physical links. More specially, it maintains good load balancing for the datacenter network using an LOBS-HC ring. In addition, due to the symmetric feature of SHM, the proposed scheme can decrease the upper bound of the average hop count of the routing paths to half of the ring size.

• Journal of KIISE:Computer Systems and Theory
• /
• v.33 no.8
• /
• pp.505-513
• /
• 2006

To satisfy SLA(Service Level Agreement) contract between a client and a service provider, the client requests are classified and processed by priorities. In providing differentiated service, a request from a client who has low priority can be dealt with less important. In this paper, we study static and dynamic partitioning mechanism of web servers and also admission control policy for multiclass request distribution. Through simulation experiments, we analyze web server throughput and response time considering SLA.

• ETRI Journal
• /
• v.42 no.5
• /
• pp.686-699
• /
• 2020

In this paper, we propose a supervised-learning-based spatial performance prediction (SLPP) framework for next-generation heterogeneous communication networks (HCNs). Adaptive asset placement, dynamic resource allocation, and load balancing are critical network functions in an HCN to ensure seamless network management and enhance service quality. Although many existing systems use measurement data to react to network performance changes, it is highly beneficial to perform accurate performance prediction for different systems to support various network functions. Recent advancements in complex statistical algorithms and computational efficiency have made machine-learning ubiquitous for accurate data-based prediction. A robust network performance prediction framework for optimizing performance and resource utilization through a linear discriminant analysis-based prediction approach has been proposed in this paper. Comparison results with different machine-learning techniques on real-world data demonstrate that SLPP provides superior accuracy and computational efficiency for both stationary and mobile user conditions.

• Proceedings of the Korean Information Science Society Conference
• /
• 2003.04a
• /
• pp.109-111
• /
• 2003

클러스터 시스템의 성능 향상을 위해서는 컴퓨팅 자원을 효과적으로 사용하여야 한다. 과거에는 전체 시스템 자원을 효과적으로 사용하기 위해 각 노드들의 부하를 균등하게 하는 방향으로 연구가 진행되어 왔으나, 부하 분산 시스템이 작업의 자원 요구 형태를 고려하여 작업을 배치하는 경우 성능을 더욱 향상시킬 수 있다. 현재까지는 이런 자원 요구 형태에 대한 선행지식을 과거 작업 실행 기록에 기반하여 유추해내는 방법을 많이 사용하였으나 이 방법은 잘못된 예측을 가져와 실행시간을 증가시킬 수 있다. 본 논문에서는 이를 해결하기 위해 유효 작업수라 불리는 새로운 노드의 부하 측정 척도를 제시한다. 유효 작업수를 이용한 부하 분산 시스템은 작업의 자원 요구 사항을 알지 못하더라도 부하 분산 과정에서 작업이 잘못 배치되어 실행시간이 증가하는 경우를 방지한다. 성능분석 결과는 과거 자료에 의한 예측을 사용하는 기존 방법에 비해 전체 실행시간의 감소로 성능이 향상되었음을 보여준다.

• The Transactions of the Korean Institute of Power Electronics
• /
• v.26 no.4
• /
• pp.256-262
• /
• 2021

Six-phase motors can be used in industrial applications, such as an electric vehicle, due to their high reliability and low current magnitude per phase. An asymmetrical PMSM with two sets of three-phase windings is a commonly used structure for six-phase motors, with each winding set demonstrating a phase difference of 30°. Although the asymmetrical PMSM presents low torque ripples, its dynamic torque response deteriorates due to coupled components in the two three-phase windings. The decoupled VSD control is applied to eliminate the coupling effect. Load ratio control of two inverters for the six-phase PMSM is proposed in this study. DQ currents are controlled on the basis of two synchronous reference frames, and the six-phase drive system can be changed to a three-phase drive system when one inverter presents fault conditions. The operation and effectiveness of the proposed algorithm is verified through simulation and experiments. The six-phase drive system is transferred to a three-phase drive system by changing the current reference of the second DQ reference frame. Moreover, control of both torque and speed exhibits satisfactory performance before and after the mode change.

• Journal of Korea Game Society
• /
• v.13 no.2
• /
• pp.39-48
• /
• 2013

In MMORPGs(Massively Multi-player Online Role-Playing Games) a large number of players actively interact with one another in a virtual world. Therefore MMORGs must be able to quickly process real-time access requests and process requests from numerous gaming users. A key challenge is that the workload of the game server increases as the number of gaming users increases. To address this workload problem, many developers apply with distributed server architectures which use dynamic map partitioning and load balancing according to the server function. Therefore most MMORPG servers partition a virtual world into zones and each zone runs on multiple game servers. These methods cause of players frequently move between game servers, which imposes high overhead for data updates. In this paper, we propose a new architecture that apply with an arithmetic server dedicated to data operation. This architecture enables the existing game servers to process more access and job requests by reducing the load. Through mathematical modeling and experimental results, we show that our scheme yields higher efficiency than the existing ones.

• The Journal of Korean Institute of Next Generation Computing
• /
• v.15 no.4
• /
• pp.18-29
• /
• 2019

A dynamic adaptive streaming server consumes high processor power because it handles a large amount of transcoding operations at a time. For this purpose, multi-processor architecture is mandatory for which effective transcoding task distribution strategies are essential. In this paper, we present the design and implementation details of the transcoding workload distribution schemes at a 2-tier (frontend node and backend node) transcoding server. For this, we implemented four schemes: 1) allocation of transcoding tasks to appropriate back-end nodes, 2) task scheduling in the back-end node and 3) the communication between front-end and back-end nodes. Experiments were conducted to compare the estimated and the actual power consumption in a real testbed to verify the efficacy of the system. It also proved that the system can reduce the load on each node to optimize the power and time used for transcoding.

• KEPCO Journal on Electric Power and Energy
• /
• v.7 no.2
• /
• pp.329-333
• /
• 2021

V2G (Vehicle to Grid) technology for an EV (Electric Vehicle) has been assumed as so promising in a near future for its useful energy resource concept but still yet to be developed around the world for specific service purposes through various R&BD projects. Basically, V2G returns power stored in vehicle at a cheaper or unused time to the grid at more expensive or highly peaked time, and is accordingly supposed to provide such roles like peak shaving or load levelling according to customer load curve, frequency regulation or ancillary reserves, and balancing power fluctuation to grid from the weather-sensitive renewable sources like wind or solar generations. However, it has recently been debated over its prominent usage as diffusing EVs and the required charging/discharging infrastructure, partially for its addition of EV ownership costs with more frequent charging/discharging events and user inconvenience with a relative long-time participation in the previously engaged V2G program. This study suggests that a Korean DR (Demand Response) service integrated V2G system especially based upon a dynamic charge/pause/discharge scheme newly proposed to ISO/IEC 15118 rev. 2 can deal with these concerns with more profitable business model, while fully making up for the additional component (ex. battery) and service costs. It also indicates that the optimum economic, environmental, and grid impacts can be simulated for this V2G-DR service particularly designed for EV aggregators (V2G service providers) by proposing a specific V2G engagement program for the mediated DR service providers and the distributed EV owners.

• Annual Conference of KIPS
• /
• 2011.04a
• /
• pp.171-174
• /
• 2011

서버 클러스터에서 부하 분산기는 사용자의 요청을 각 서버로 분산시키는 역할을 한다. 리눅스 가상 서버(LVS: Linux Virtual Server)는 소프트웨어적으로 사용되는 부하 분산기로서 여러 가지 스케줄링 방식들을 가지고 있다. 그러나 부하 분산 시에 서버의 유동적인 부하 정보를 반영하지 못하는 단점이 있다. 이에 개선된 방식으로 서버의 동시 연결 개수에 따라 상한계(Upper Bound)와 하한계(Lower Bound)를 설정하고, 요청을 분산하는 동적 스케줄링(Dynamic Scheduling)이 존재한다. 그러나 서버의 상태에 따라 상한계와 하한계가 바뀔 수 있음에도 불구하고 이 값들이 고정되어 있다는 단점을 가진다. 본 논문에서는 기존 부하 분산 방법의 단점을 극복하는 서버 전력 정보에 기반한 스케줄링 방식을 제안한다. 제안된 방식은 서버의 부하 정보를 기반으로 에너지를 추정하고 전력 수치를 기반으로 LVS의 가중치 테이블을 주기적으로 갱신한다. 그리고 부하 분산기는 클라이언트로부터 요청 받은 트래픽을 각 서버의 에너지 소모 상태에 따라 적용시킴으로써 에너지 소모가 최소화되도록 부하를 분산한다. 또한 서버의 상태에 따라 상한계와 하한계가 바뀔수 있음을 고려하여 상한계와 하한계를 설정하지 않고 서버의 상태에 따라 적절하게 요청이 분배되도록 하였다. 15대의 PC를 사용하여 실험을 수행하였으며, 실험 결과는 기존 부하 분산 알고리즘 중 성능이 가장 좋은 알고리즘에 비해 서버의 성능이 동일한 경우 성능 및 소비전력 면에서 거의 동등하였고, 서버의 성능이 상이한 경우 50.2% 성능 향상 및 27.3% 소비 전력 절감을 확인하였다.

• KIPS Transactions on Computer and Communication Systems
• /
• v.2 no.7
• /
• pp.283-292
• /
• 2013

In order to ensure high performance, all the servers in an existing server cluster are always On regardless of number of real-time requests. They ensure QoS, but waste server power if some of them are idle. To save energy consumed by servers, the server power mode control was developed by shutdowning a server when a server is not needed. There are two types of server power mode control depending on when a server is actually turned off if the server is selected to be off: static or dynamic. In a static mode, the server power is actually turned off after a fixed time delay from the time of the server selection. In a dynamic mode, server power is actually turned off if all the services served in the server are done. This corresponds to a turn off after a variable time delay. The static mdoe has disadvantages. It takes much time to find an optimal shutdown time manually through repeated experiments. In this paper, we propose a dynamic shutdown method to overcome the disadvantages of static shutdown. The proposed method allows to guarantee user QoS with good power-saving because it automatically approaches an optimal shutdown time. We performed experiments using 30 PCs cluster. Experimental results show that the proposed dynamic shutdown method is almost same as the best static shutdown in terms of power saving, but better than the best static shutdown in terms of QoS.