• The KIPS Transactions:PartA
• /
• v.8A no.4
• /
• pp.299-310
• /
• 2001

In this paper, we study the methodology of distributing the requests of clients based on the state of real server in the LVS(Linux virtual server) clustering system. The WLC(weighted least connection) algorithm in the LVS cluster system is studied. The load distributing algorithm which assigns a weight into each real server is devised by considering the load of real servers. The load test is executed to estimate the load of real servers using a load generating tool. Throughout the result of the experiment, we suggest new load distributing algorithm based on the usage of physical memory of each real server. It is shown that the correction potentiality of new algorithm is somewhat better than that of the WLC algorithm.

• The KIPS Transactions:PartA
• /
• v.14A no.1 s.105
• /
• pp.63-72
• /
• 2007

This paper introduces a load-balancing algorithm focusing on distributing web requests evenly into the web cluster servers. The load-balancing algorithms based on conventional periodic load-information update mechanism are not suitable for dynamic page applications, which are common in Cyber Foraging services, due to the problems caused by periodic synchronized load-information updating and the difficulties of work load estimation caused by embedded executing scripts of dynamic pages. Update-on-Finish algorithm solves this problem by using non-periodic load-update mechanism, and the web switch knows the servers' real load information only after their reporting and then distributes new loads according to the new load-information table, however it results in much communication overhead. Our proposed mechanism improve update-on-finish algorithm by using K-Percents-Finish mechanism and thus largely reduce the communication overhead. Furthermore, we consider the different capabilities of servers with a threshold Ti value and propose a load-balancing algorithm for servers with various capabilities. Simulation results show that the proposed K-Percents-Finish Reporting mechanism can at least reduce 50% communication overhead than update-on-finish approach while sustaining better load balancing performance than periodic mechanisms in related work.

• Proceedings of the Korean Information Science Society Conference
• /
• 2004.04a
• /
• pp.412-414
• /
• 2004

수십 대의 PC들로 구성된 학교 PC 실 또는 교육 목적 PC 실에서는 컴퓨터들이 분산 구조로 되어 있어서 각 컴퓨터별로 셋업. 유지보수, 업그레이드가 각각 따로따로 수행된다. 이러한 분산 구조에 대한 대안으로 씬 클라이언트 컴퓨팅 환경을 고려해 볼 수 있다. 씬 클라이언트 컴퓨팅 환경에서, 클라이언트 쪽 장치는 사용자에게 친숙한 GUI 와 멀티미디어 지원과 함께 주로 IO 기능들을 제공하는 반면에 터미널 서버라 불리는 원격 서버들은 컴퓨팅 파워를 제공한다. 이 환경에서는 많은 클라이언트를 지원하기 위해서 터미널 서버들을 클러스터로 구성할 수 있다. 그러나 이러한 구조에서는 터미널 세션의 유지와 사용자의 다양한 컴퓨팅 사용 패턴 요인으로 부하 분산이 어렵고 결과적으로 터미널 서버 자원의 활용도가 낮아지는 단점을 가진다. 이러한 단점을 보완하기 일해 본 논문에서는 적응적 터미널 클러스터를 제안한다. 이 구조에서는 부하가 적은 그룹에 속한 터미널 서버가 부하가 큰 그룹으로 실시간에 동적으로 재 할당될 수 있다. 제안된 적응적 터미널 클러스터를 일반적인 터미널 클러스터와 그룹 기반 비적응적 터미널 클러스터와 비교하고 실험을 통해 제안된 방법의 유효성을 검증하였다.

• Journal of KIISE:Computer Systems and Theory
• /
• v.37 no.4
• /
• pp.195-204
• /
• 2010

Due to improved wireless communication technologies, it is possible to provide multimedia streaming service for mobile device clients like PDAs and cellphones. Wireless networks are serviced on low bandwidth channels and mobile devices work on limited hardware specifications. In these conditions, transcoding technologies are needed to adapt the media for streaming services to given mobile environments. To transcode from the source media to the target media for corresponding grades, transcoding servers perform transcoding jobs as exhausting their resources. Since various transcoding loads occur according to the target transconding grades, an effective transcoding load balancing policy is required among transcoding servers. In addition to transcoding process, servers should maintain QoS streams for mobile clients for total serviced times. It requires real-time requirements to support QoS for various mobile clients. In this paper, a new transcoding load distribution method is proposed. The proposed method can be driven for fair load balance between distributed transcoding servers. Based on estimated transcoding time, movie information and target transcoding bit-rate, it provides fair transcoding load distribution and also performs admission control to support QoS streams for mobile clients.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.18 no.6
• /
• pp.1087-1094
• /
• 2023

Due to the development of technologies, such as big data, cloud, IoT, and AI, The high data throughput is required, and the importance of network flexibility and scalability is increasing. However, existing network systems are dependent on vendors and equipment, and thus have limitations in meeting the foregoing needs. Accordingly, SDN technology that can configure a software-centered flexible network is attracting attention. In particular, a load balancing method based on SDN can efficiently process massive traffic and optimize network performance. In the existing load balancing studies in SDN environment have limitation in that unnecessary traffic occurs between servers and controllers or performing load balancing only after the server reaches an overload state. In order to solve this problem, this paper proposes a method that minimizes unnecessary traffic and appropriate load balancing can be performed before the server becomes overloaded through a method of assigning weights to servers in stages according to server load.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.18 no.9
• /
• pp.2169-2175
• /
• 2014

In this paper we have proposed a new distributed architecture of session-based instant message system, in which multiple servers can distribute loads effectively to improve system extendability. The server of session-based instant message system should manage SIP sessions to transfer instant message with multimedia data to each users, and it should process large amount of presence informations from each users. Therefore load of the server should be distributed as the number of users increases. In our system, when server load increases according to large user numbers, the load are distributed by selecting a server with the lowest load value, or new server is created to allocate the message processing. Therefore the overall load of instant message system can be dynamically distributed. A new presence information data format with added elements for controlling server load has been designed for this purpose, and SIP messages exchanging procedure between servers and users has been also suggested. The performance of our proposed system has been analysed by simulation, and showed improvement of performance.

• Journal of Korea Game Society
• /
• v.12 no.3
• /
• pp.69-76
• /
• 2012

At MMORPGs that has the numerous game users with many interactions at the same virtual world, massive loading that results in delays, resource shortages, and other such problems occur. To solve this, many developers devote research to load-balancing servers, yet due to steady and dynamic map divisions, such research is unreliable. This paper proposes a game user-oriented load balancing scheme for the load balancing of MMORPG servers. This scheme shows effectiveness at dealing with hot-spots and other gatherings of game users at specific servers compared to previous methods.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.16 no.5
• /
• pp.817-824
• /
• 2021

The SDN(Software Defined Networking) technology, which appeared to overcome the limitations of the existing network system, resolves the rigidity of the existing system through the separation of HW and SW in network equipment. These characteristics of SDN provide wide scalability beyond hardware-oriented network equipment, and provide flexible load balancing policies in data centers of various sizes. In the meantime, many studies have been conducted to apply the advantages of SDN to data centers and have shown their effectiveness. The method mainly used in previous studies was to periodically check the server load and perform load balancing based on this. In this method, the more the number of servers and the shorter the server load check cycle, the more traffic increases. In this paper, we propose a new load balancing technique that can eliminate unnecessary traffic and manage server resources more efficiently by reporting to the controller when a specific level of load occurs in the server to solve this limitation.

• Journal of the Korea Society of Computer and Information
• /
• v.13 no.1
• /
• pp.55-64
• /
• 2008

Recently virtualization becomes one of the most popular research topics. and a lot of software products which is related to virtualization have been released accordingly. Server virtualization, which virtualize physical servers to supply many virtual servers, provides very efficient way to build network-based servers. In this paper, we design and implement virtual desktop service which is based on server virtualization. We also propose load balancing scheme for the virtual desktop service. The proposed virtual desktop service and its load balancing scheme provides a cost-effective way to build high-performance remote desktop service.

• Proceedings of the Korean Information Science Society Conference
• /
• 2001.04b
• /
• pp.211-213
• /
• 2001

공간 데이터베이스 시스템에서는 데이터의 보안과 사용자의 편의성을 제공하기 위해 사용자가 원하는 공간데이터만으로 구성된 공간 뷰를 제공한다. 클라이언트/서버 환경의 공간 데이터베이스 시스템에서 다수의 클라이언트에 의해 공간 뷰에 대한 질의가 요청 될 시 대용량의 데이터를 처리하기 위한 서버의 I/O 연산의 수행비용과 질의처리 비용 및 결과 데이터의 전송을 위한 전송 비용이 서버의 부하를 일으키고 질의 처리속도의 저하를 야기시킨다. 본 논문에서는 클라이언트/서버 환경의 공간 데이터베이스 시스템에서 공간 뷰의 생성 과정을 서버와 클라이언트에 분산시킨 크라이언트 실체화 기법을 제안한다. 공간 뷰 생성의 질의처리를 서버와 클라이언트에 분산시켜 대용량의 데이터와 복잡한 공간 연산에 따른 공간 뷰 생성과정의 서버 부하를 감소시키고 클라이언트에 실체화 함으로 해서 공간뷰에 대한 질의처리 요구에 따른 서버의 병목현상과 서버 부하를 감소시켜 사용자 응답시간을 최소화한다.