• Proceedings of the Korean Information Science Society Conference
• /
• 1998.10a
• /
• pp.30-32
• /
• 1998

멀티프로세서 시스템을 설계할 경우에는 공유메모리 구조와 메시지 전달방법의 두 가지의 패러다임을 바탕으로 하게 된다. 데이터 분할과 동적 부하 분산 문제를 단순화시틸 수 있으며 확장성을 용이하게 지원하는 장점을 가지고 있는 공유메모리 구조의 멀티프로세서 시스템에서 각 프로세서가 자신의 전용 캐시를 가지는 경우에는 메인 메모리와 이러한 전용 캐시내에 존재하는 데이터사본간에 일관성 문제가 발생한다. 본 논문에서는 일관성 유지를 위해 제안되어 있는 여러 알고리즘 중 처리 노드와 고대역 저지연 인터커넥션 네트워크로 구성되는 공유메모리 구조의 멀티프로세서 프로토타입인 DASH 프로토콜을 지원하기 위한 전용 버스 시스템을 완전 개방형인 IEEE Futurebus+ 스탠다드에 준비하여 설계한 다음, 이 시스템이 DASH 프로토콜을 지원하려 캐시의 일관성을 유지하기 위해 필요한 각종 행동과 기존의 범용 버스 시스템이 수행하는 행동의 병렬 처리를 지원할 수 있음을 시뮬레이션으로 증명한다.

• 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.

• KIISE Transactions on Computing Practices
• /
• v.20 no.9
• /
• pp.507-512
• /
• 2014

Recently, Distributed computing processing begins using both CPU(Central processing unit) and GPU(Graphic processing unit) to improve the performance to overcome darksilicon problem which cannot use all of the transistors because of the electric power limitation. There is an integrated graphics processor that CPU and GPU share memory and Last level cache(LLC). But, There is no LLC access rules between CPU and GPU, so if GPU and CPU processes run together at the same time, performance of both processes gets worse because of the contention on the LLC. This Paper gives evidence to prove the need of the Cache Partitioning and is mentioned about the cache partitioning design using page coloring to allocate the L3 Cache space only for the GPU process to guarantee GPU process performance.

• Journal of the Korea Society of Computer and Information
• /
• v.14 no.4
• /
• pp.9-17
• /
• 2009

In this paper, we propose a dynamic load distribution technique at the operating system level in mobile storage systems with a heterogeneous storage pair of a small form-factor and disk and a flash memory, which aims at saving energy consumption as well as enhancing I/O performance. Our proposed technique takes a combinatory approach of file placement and buffer cache management techniques to find how the load can be distributed in an energy and performance-aware way for a heterogeneous mobile storage air of a hard disk and a flash memory. We demonstrate that the proposed technique provides better experimental results with heterogeneous mobile storage devices compared with the existing techniques through extensive simulations.

• KIISE Transactions on Computing Practices
• /
• v.22 no.1
• /
• pp.44-49
• /
• 2016

Recently, data has been growing dramatically along with the growth of social media and digital devices. A distributed memory processing system has been used to efficiently process large amounts of data. However, if a load is concentrated in a certain node in distributed environments, a node performance significantly degrades. In this paper, we propose a load balancing scheme to distribute load in a distributed memory environment. The proposed scheme replicates hot data to multiple nodes for managing a node's load and migrates the data by considering the load of the nodes when nodes are added or removed. The client reduces the number of accesses to the central server by directly accessing the data node through the metadata information of the hot data. In order to show the superiority of the proposed scheme, we compare it with the existing load balancing scheme through performance evaluation.

• Proceedings of the Korean Information Science Society Conference
• /
• 2001.10c
• /
• pp.697-699
• /
• 2001

가상환경에서의 실시간 상호작용을 지원하기 위해 가상세계의 데이터를 서버에서 클라이언트로 복제 해놓는 방법이 쓰여지고 있다. 효과적인 복제를 위해서는 두 가지 기법 - 우선 순위기반의 객체 전송과 캐싱 및 프리패칭 기법이 쓰이고 있다. 기존의 캐시와 프리패칭 기법들은 사용자와 객체간의 지역적 관계를 이용하고 있지만, 이는 어떤 종류의 객체가 사용자에게 더 중요한 지를 결정할 수 없는 단점이 있다. 본 연구에서는 지역적 관계 뿐 아니라 개별 사용자의 객체에 대한 관심도로부터 발생하는 객체에 대한 우선 순위를 이용하여 사용자 기반의 캐싱과 프리패칭을 사용하는 분산 가상환경 하에서의 확장성 있는 데이터 관리 기법을 제안한다. 또한 이 기법에서는 사용자의 행동 모드를 구분함으로써 더욱 높은 캐시 적중 비율을 얻을 수 있다.

• Proceedings of the Korea Contents Association Conference
• /
• 2003.11a
• /
• pp.234-244
• /
• 2003

In distributed file-systems, cooperative caching algorithm which owns the data cached at each node jointly is used to reduce an expense of disk access. Cooperative caching algorithm is the method that increases a cache hit-ratio and decrease a disk access as it holds the cache information of distributed systems in common and makes cache larger virtually. Recently, several cooperative caching algorithms decrease the message costs by using approximate information of the cache and increase the cache hit-ratio by using local and global cache fields dynamically. And they have an advantage that increases the whole field hit-ratio by sending a replaced block to the idel node on cache replacement in order to maintain the replaced block in the cache field. However the wrong approximate information deteriorates the performance, the concistency maintenance goes to great expense to exchange messeges and the cost that manages Age-information of each node to choose the idle node increases. In this thesis, we propose a cooperative cache algorithm that maintains correct cache information, minimizes the maintance cost for consistency and the management cost for cache Age-information. Also, we show the superiority of our algorithm through the performance evaluation.

• Journal of KIISE:Computing Practices and Letters
• /
• v.16 no.4
• /
• pp.472-476
• /
• 2010

Nowadays, as the indices of the major search engines grow to a tremendous proportion, vertical search services can help customers to find what they need. Real time search is valuable because it lets you know what's happening right now on any given topic. In this paper, we designed a new architecture to implement a high performance real time search system. Based on the real time search's characters, we divided the whole system to two parts which are collection system and search system. The evaluation results showed that our design has the potential to provide the real time search transparent scalability while maintaining the replication overhead costs in check.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2015.04a
• /
• pp.676-679
• /
• 2015

빅데이터 분석을 위한 Hadoop 환경에서 고성능 저장장치인 SSD에 대한 중요성이 증가하면서 일반적으로 사용되는 저장장치인 HDD와 혼합하여 사용하는 연구들이 주목 받고 있다. 특히 SSD를 HDD의 캐시로 사용했을 때 저장장치에 대한 I/O 성능을 향상할 수 있다는 연구 결과들이 있다. 본 연구는 이를 바탕으로 SSD를 HDD의 캐시로 사용한다. HDFS는 저장장치에 접근하여 I/O를 수행하는데 기존에는 로컬 서버에서 캐시 미스가 발생한 경우 로컬 HDD로 접근한다. 이러한 방식은 접근하는 데이터에 따라 SSD의 높은 Bandwidth를 활용하지 못하게 되는 경우를 발생시키고 그 결과 특정 서버의 I/O 지연으로 전체 분산 처리의 성능을 저하시킬 수 있다. 이를 해결하기 위해 본 연구는 HDFS 레벨에서 로컬 서버의 HDD와 데이터 복제본들이 저장된 원격 서버의 SSD에서 I/O를 수행하는 경우에 대해 수식을 통해 비용을 비교한다. 그 결과 항상 기대 성능이 높은 저장 장치를 선택하여 데이터를 읽어오게 함으로써 기존 방식보다 성능이 개선될 수 있음을 입증한다.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.18 no.6
• /
• pp.1095-1102
• /
• 2023

Content Centric Network (CCN) is a network that emerged to improve the existing network structure and communicates based on content names instead of addresses. It utilises caches to distribute traffic and reduce response time by delivering content from intermediate nodes. In this paper, we propose a popularity-based caching policy to efficiently utilise the limited CS space in CCN environment. The performance of CCNs can vary significantly depending on which content is prioritised to be stored and released. To achieve the most efficient cache replacement, we propose a real-time content popularity-based efficient cache replacement policy that calculates and prioritises content popularity based on constructor popularity, constructor distance, and content hits, and demonstrate the effectiveness of the new policy through experiments.