• Journal of Korea Multimedia Society
• /
• v.14 no.6
• /
• pp.835-848
• /
• 2011

In this paper, we propose a client server based distributed/parallel system to effectively visualize huge geographic data. The system consists of a web-based client GUI program and a distributed/parallel server program which runs on multiple PC clusters. To make the client program run on mobile devices as well as PCs, the graphical user interface has been designed by using JOGL, the java-based OpenGL graphics library, and sending the information about current available memory space and maximum display resolution the server can minimize the amount of tasks. PC clusters used to play the role of the server access requested geographic data from distributed disks, and properly re-sample them, then send the results back to the client. To minimize the latency happened in repeatedly access the distributed stored geography data, cache data structures have been maintained in both every nodes of the server and the client.

• Journal of the Korea Computer Industry Society
• /
• v.3 no.1
• /
• pp.77-84
• /
• 2002

An cluster system used consist of large number of running servers, one has the problem that does the low availability occured by the high chance of the server failures and it is difficult to provide occuring software aging. In this paper, running cluster web servers consists of n primary servers and k backup servers, based on the operational parameters such as number of running primary servers, number of backup severs, rejuvenation period, rejuvenation time, failure rate of sewers, repair rate of servers, unstable rate of servers. We calculate to evaluate the rejuvenation policy such steady-state probabilities, downtime, availability, and downtime cost. We validate the solutions of mathematical model by experiments based on various operation parameters and find that the software rejuvenation method can be adopted as prventive fault tolerant technique for stability of system. The failure rate and unstable rate of the servers are essential factors for decision making of the rejuvenation policies.

• Journal of KIISE:Computer Systems and Theory
• /
• v.35 no.3
• /
• pp.111-119
• /
• 2008

This paper combines fuzzy concepts with an existing dynamic performance isolation technique in order to improve the response time performance of a Web server supporting differentiated services. A load balancing mechanism based on the fuzzy control technique is developed in such a way that ambiguous situations caused by workload estimation of cluster-based Web servers, client request rates, and dynamic request rates can be represented in an effective way. In addition, we verify that the fuzzy-based performance isolation technique improves the performance and robustness of differentiated service systems efficiently through comparing 95-percentile of response time between the fuzzy-based Performance isolation technique and the existing one, which do not use the fuzzy concept.

• Journal of KIISE:Computer Systems and Theory
• /
• v.32 no.11_12
• /
• pp.717-721
• /
• 2005

With HTTP/1.0, a single request means a single HTTP connection so that the granular unit of dispatching is the same as real load. But with persistent HTTP connection, multiple requests may arrive on a single TCP connection. Therefore, a scheme that dispatches load at the granularity of individual requests constrains the feasible dispatching policies In this paper we propose a new connection dispatching polity for supporting HTTP/1.1 persistent connections in cluster-based Web servers. When the request of a base html file arrives, the dispatcher gets the subsequent load arriving on that connection using the embedded objects information. After the dispatcher stores the load information in Load Table, the dispatcher employs the connection aging strategy on live persistent connections on the passage of time. The results of simulation show about $1.7\%\~16.8\%$ improved average response time compared to existing WLC algorithm.

• The KIPS Transactions:PartA
• /
• v.13A no.7 s.104
• /
• pp.615-622
• /
• 2006

Caching in a Wireless Internet Proxy Server Cluster Environment has an effect that minimizes the time on the request and response of Internet traffic and Web user As a way to increase the hit ratio of cache, we can use a hash function to make the same request URLs to be assigned to the same cache server. The disadvantage of the hashing scheme is that client requests cannot be well-distributed to all cache servers so that the performance of the whole system can depend on only a few busy servers. In this paper, we propose an improved load balancing scheme using hashing and Round Robin scheme that distributes client requests evenly to cache servers. In the existing hashing scheme, if a hashing value for a request URL is calculated, the server number is statically fixed at compile time while in the proposed scheme it is dynamically fixed at run time using round robin method. We implemented the proposed scheme in a Wireless Internet Proxy Server Cluster Environment and performed experiments using 16 PCs. Experimental results show the even distribution of client requests and the 52% to 112% performance improvement compared to the existing hashing method.

• Journal of KIISE:Computing Practices and Letters
• /
• v.14 no.7
• /
• pp.669-679
• /
• 2008

With the increasing demands of information sharing and searches via the web, the web search engine has drawn much attention. Although many researches have been done to solve technical challenges to build the web search engine, the issue regarding its query processing system is rarely dealt with. Since the software architecture and operational schemes of the query processing system are hard to elaborate, we here present related techniques implemented on a commercial system. The implemented system is a very large-scale system that can process 5-million user queries per day by using index files built on about 65-million web pages. We implement a multi-level cache scheme to save already returned query results for performance considerations, and the multi-level cache is managed in 4-level cache storage areas. Using the multi-level cache, we can improve the system throughput by a factor of 4, thereby reducing around 70% of the server cost.

• Genomics & Informatics
• /
• v.6 no.3
• /
• pp.157-159
• /
• 2008

Multiple sequence alignment (MSA) is the most important step for many of biological sequence analyses, homology search, and protein structural assignments. However, large amount of data make biologists difficult to perform MSA analyses and it requires much computational time to align many sequences. Here, we have developed a simple and fast web alignment tool for aligning, editing, and visualizing large amount of sequence data. We used a cluster server installed ClustalW-MPI using web services and message passing interface (MPI). It also enables users to edit multiple sequence alignments for manual editing and to download the input data and results such as alignments and phylogenetic tree.

• Proceedings of the Korean Information Science Society Conference
• /
• 2002.04a
• /
• pp.607-609
• /
• 2002

일반적으로 부하 분산 정책은 사용자의 서비스 요청 특성에 관계없이 들어오는 그대로 처리 하고 있다. 본 논문에서는 이러한 것을 개선하기 위하여 차별 큐를 가진 최단 작업 우선 기 법을 제안하였다. 제안한 기법은 부하 분산 서버에 사용자의 서비스 요청이 들어오면 해당 요청의 처리시간을 파악하고 이미 들어온 서비스 요청들과 비교해서 먼저 처리가 완료될 수 있는 서비스 요청을 부하 분산 정책에 할당해주는 전략을 사용함으로써, 사용자들의 평균 응답 시간을 향상시키도록 하였다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2002.10e
• /
• pp.532-534
• /
• 2002

본 논문에서는 UDP／IP기반 SAN의 구현을 소개하였고 TCP／IP기반 SAN 및 로컬디스크와의 성능 비교를 통해 UDP／IP기반 SAN이 클러스터 웹서버에서 사용될 수 있는 가능성을 보였다. 실험결과 UDP／IP기반 SAN은 TCP／IP기반 SAN의 경우보다 약 20%정도 성능이 우수함을 볼 수 있다. 이는 UDP의 경우 TCP의 프로토콜 오버헤드가 없고 slow start의 영향을 받지 않으며 ACK를 기다릴 필요가 없기 때문이다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2003.04d
• /
• pp.563-565
• /
• 2003

최근 클러스터에 대한 연구는 내용 기반 클러스터의 부하 분배와 캐시 정책에 집중되고 있다. 본 논문에서는 웹 서비스의 고가용성 및 확장성을 제공하는 클러스터 환경에서 힌트 기반 협력적 캐싱의 캐시 적중률을 향상시키기 위해 기존의 DFR 기법을 개선하였다. 서비스 접근 확률을 이용하여 주 복사본과 종속 복사본을 선택적으로 제거하는 메모리 교체 방법을 제시하였으며, DFR 방식과 성능을 비교, 분석한 결과 DFR 방식보다 적은 디스크 접근률을 얻을 수 있었다.