• 정보처리학회논문지A
• /
• 제14A권1호
• /
• pp.63-72
• /
• 2007

본 논문에서 웹 서버 클러스터 환경에서 웹 요청들의 분산을 위한 부하 분산 기법을 제안한다. 전통적인 주기적 부하 정보 갱신 기반의 부하 분산 기법에서는 동기화된 부하 정보와 동적 페이지에 포함된 스크립트들의 갱신 정보 때문에 Cyber Foraging 서비스 같은 동적 웹 응용프로그램에서는 적합하지 않다. 이를 해결하기 위해 Update-on-Finish 부하 분산 알고리즘은 비주기적인 부하 갱신 방법을 사용하고 있다. 웹 스위치는 비주기적인 부하 리포팅 후, 서버들의 실제 부하 정보를 알 수 있다. 그런후 실제 정보에 따라 부하 분산 스케줄을 재배열한다. 하지만 Update-on-Finish 알고리즘의 경우 부하 정보를 유지하기 위한 통신 부하가 크다는 문제점을 가진다. 본 논문에서는 각 서버가 K%의 작업을 마친 후 비주기적 부하 정보 보고를 통하여 부하 감소시킨 방법을 제안한다. 또한 서버의 처리능력이 다른 환경을 고려하여 서로 다른 threshold Ti값을 적용함으로써, 다양한 처리 능력을 가진 서버들을 위한 로드 밸런싱 알고리즘으로 확장하여 제안하고 있다. 시뮬레이션 결과에서 제안된 K-Percent-Finish Reporting 방법은 Update-on-Finish 방법보다 최소 50% 이상의 통신 부하를 감소시키면서, 기존 주기적 부하정보 갱신 기반의 관련 연구들보다 향상된 시스템 처리 능력을 보여주고 있다.

• ETRI Journal
• /
• 제36권5호
• /
• pp.761-771
• /
• 2014

Long-Term Evolution employs a hard handover procedure. To reduce the interruption of data flow, downlink data is forwarded from the serving eNodeB (eNB) to the target eNB during handover. In cellular networks, unbalanced loads may lead to congestion in both the radio network and the backhaul network, resulting in bad end-to-end performance as well as causing unfairness among the users sharing the bottleneck link. This work focuses on congestion in the transport network. Handovers toward less loaded cells can help redistribute the load of the bottleneck link; such a mechanism is known as load balancing. The results show that the introduction of such a handover mechanism into the simulation environment positively influences the system performance. This is because terminals spend more time in the cell; hence, a better reception is offered. The utilization of load balancing can be used to further improve the performance of cellular systems that are experiencing congestion on a bottleneck link due to an uneven load.

• International Journal of Computer Science & Network Security
• /
• 제21권2호
• /
• pp.158-170
• /
• 2021

Virtualization technologies are being adopted and broadly utilized in many fields and at different levels. In cloud computing, achieving load balancing across large distributed virtual machines is considered a complex optimization problem with an essential importance in cloud computing systems and data centers as the overloading or underloading of tasks on VMs may cause multiple issues in the cloud system like longer execution time, machine failure, high power consumption, etc. Therefore, load balancing mechanism is an important aspect in cloud computing that assist in overcoming different performance issues. In this research, we propose a new approach that combines the advantages of different task allocation algorithms like Round robin algorithm, and Random allocation with different threshold techniques like the VM utilization and the number of allocation counts using least connection mechanism. We performed extensive simulations and experiments that augment different scheduling policies to overcome the resource utilization problem without compromising other performance measures like makespan and execution time of the tasks. The proposed system provided better results compared to the original round robin as it takes into consideration the dynamic state of the system.

• Journal of Electrical Engineering and Technology
• /
• 제12권6호
• /
• pp.2227-2236
• /
• 2017

In order to solve the neutral-point voltage fluctuation problem of three-phase three-level T-type inverters (TPTLTIs), the unbalance characteristics of capacitor voltages under different switching states and the mechanism of neutral-point voltage fluctuation are revealed. Based on the mathematical model of a TPTLTI, a feed-forward voltage balancing control strategy of DC-link capacitor voltages error is proposed. The strategy generates a DC bias voltage using a capacitor voltage loop with a proportional integral (PI) controller. The proposed strategy can suppress the neutral-point voltage fluctuation effectively and improve the quality of output currents. The correctness of the theoretical analysis is verified through simulations. An experimental prototype of a TPTLTI based on Digital Signal Processor (DSP) is built. The feasibility and effectiveness of the proposed strategy is verified through experiment. The results from simulations and experiment match very well.

• 정보과학회 컴퓨팅의 실제 논문지
• /
• 제22권1호
• /
• pp.26-31
• /
• 2016

Migrim(Migration enhanced Grid Middleware)은 임베디드 단말과 다중 서버들 간의 통신 미들웨어이다. 기존의 통신 미들웨어는 단말이 서버의 자원 상태를 알지 못하기 때문에, 특정 서버에 부하가 집중될 경우에도 그 서버로 계속 요청을 보내는 문제가 있다. 또한 단말은 서버가 가진 정보를 알 수 없기 때문에, 트랜잭션 처리 결과에 대한 정확도를 보장하기 어렵다. 병목현상과 응답 정확도 문제 해결을 위해, 본 논문에서는 Migrim에 적용한 커넥션 이전과 트랜잭션 이전 기법을 제안한다. 커넥션을 다른 서버로 위임하는 과정인 커넥션 이전을 통해 전체 서버에 부하를 균등하게 분산시킬 수 있고, 트랜잭션 처리를 다른 서버로 위임하는 과정인 트랜잭션 이전을 통해 응답의 정확도를 높일 수 있다.

• ETRI Journal
• /
• 제41권2호
• /
• pp.197-206
• /
• 2019

Software-defined networking (SDN) is a modern approach for current computer and data networks. The increase in the number of business websites has resulted in an exponential growth in web traffic. To cope with the increased demands, multiple web servers with a front-end load balancer are widely used by organizations and businesses as a viable solution to improve the performance. In this paper, we propose a load-balancing mechanism for SDN. Our approach allocates web requests to each server according to its response time and the traffic volume of the corresponding switch port. The centralized SDN controller periodically collects this information to maintain an up-to-date view of the load distribution among the servers, and incoming user requests are redirected to the most appropriate server. The simulation results confirm the superiority of our approach compared to several other techniques. Compared to LBBSRT, round robin, and random selection methods, our mechanism improves the average response time by 19.58%, 33.94%, and 57.41%, respectively. Furthermore, the average improvement of throughput in comparison with these algorithms is 16.52%, 29.72%, and 58.27%, respectively.

• Journal of Power Electronics
• /
• 제10권5호
• /
• pp.572-578
• /
• 2010

This paper discusses a transformerless shunt static compensator (STATCOM) based on a modular multilevel converter (MMC). It introduces a new time-discrete appropriate current control algorithm and a phase-shifted carrier modulation strategy for fast compensation of the reactive power and harmonics, and also for the balancing of the three-phase source side currents. Analytical formulas are derived to demonstrate the accurate mechanism of the stored energy balancing inside the MMC. Various simulated waveforms verify that the MMC based STATCOM is capable of reactive power compensation, harmonic cancellation, and simultaneous load balancing, while controlling and balancing all of the DC mean voltages even during the transient states.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2008년도 하계종합학술대회
• /
• pp.249-250
• /
• 2008

In the field of mobile communications, load balancing between 3GPP LTE and other networks is expected to be an important topic from the perspective of current technical development. The load balancing technique includes the definition of functions and structures needed to incooperate 3G LTE E-Node B and base-stations designed for heterogeneous networks. The major goal attained from this paper is to ensure an algorithm for the mechanism of load balancing and to achieve a technical leading for the next generation mobile network.

• 제어로봇시스템학회논문지
• /
• 제19권6호
• /
• pp.501-507
• /
• 2013

This paper presents the development and balancing control of GYROBO, a one wheeled mobile robot system. GYROBO is a disc type one wheel mobile robot that has three actuators, a drive motor, a spin motor, and a tilt motor. The dynamics and kinematics of GYROBO are analyzed, and simulation studies conducted. A one-wheeled robot, GYROBO is built and its balancing control is performed. Experimental studies of GYROBO's balancing abilities are conducted to demonstrate the gyroscopic effects generated by the spin and tilt angles of a flywheel.

• 한국통신학회논문지
• /
• 제33권7A호
• /
• pp.757-767
• /
• 2008

차세대 이동 통신망의 발전으로 상용 망과 이 기종 망이 서로 겹쳐진 다중 망 구조가 형성될 것이다. 이러한 여러 망이 겹쳐진 다중 망 구조에서 전체 무선자원의 사용 효율을 높이고 Eb/Io (bit energy-to-interference plus noise density ratio)를 개선하기 위해 이기종망간 혹은 동종망간 상호 연동 기술이 필요하다. 본 논문에서는 이러한 동종망과 이기종망들이 서로 겹쳐져 있는 다중 셀 환경에서 load를 효율적으로 분산시킴으로써 무선 자원의 효율을 높이고 신호의 간섭양을 줄여 사용자의 QoS 보장 및 성능 개선을 위한 soft load balancing algorithm을 제안 및 분석한다. 본 논문에서는 수학적 분석을 통해 효율적인 load sharing을 위한 최적의 LBR (load balancing ratio) 을 구하고 이를 이용하여 각 network에 traffic 양을 할당하여 서비스한다. 그러므로 soft load balancing 알고리듬은 보다 안정적인 데이터 전송을 가능하게 한다.