• Proceedings of the Korean Information Science Society Conference
• /
• 1999.10c
• /
• pp.244-246
• /
• 1999

본 논문에서는, multiple-buffered crossbar 스위치를 이용한 다층 연결 망의 성능 분석 모형을 제안하고, 스위치에 장착된 buffer의 개수 증가에 따른 성능 향상 추이를 분석하였다. Buffered 스위치 기법은 다층 연결 망 (Multistage Interconnection Network, MIN)의 내부의 데이터 충돌 문제를 효과적으로 해결할 수 있는 방법으로 알려져 있다. 제안된 성능 분석 모형은 먼저 네트웍 내부 임의 스위치 입력 단에 유입되는 데이터 패킷이 buffered 스위치 내부에서 전송되는 패턴을 확률적으로 분석하여 수립하였다. 분석 모형의 수학적 복잡도 절감을 위하여 확률식 유도 과정에 정상상태 확률 rosa(steady state probability)을 도입하였다. 제안한 모형은 스위치의 크기 및 스위치에 장착된 buffer의 수와 무관하게 확대 적용이 가능하다. 제안한 수학적 성능 분석 연구의 실효성 검증을 위하여 병행된 시뮬레이션 처리 결과는 상호 미세한 오차 범위 내에서 모형의 예측 데이터와 일치하는 결과를 보여 분석 모형의 타당성을 입증하였다. 2$\times$2 스위치로 구성된 8$\times$8 MIN을 대상으로 분석을 시행한 결과 스위치에 2~4개의 buffer를 장착했을 경우 unbuffer 스위치 경우와 비교하여 네트웍 정상상태 Throughout의 증가율이 높고 네트웍 Delay 또한 낮아져 효율적인 것으로 나타났다. 따라서 2~2 crossbar 스위치로 구성된 MIN의 경우 스위치에 장착된 buffer의 개수가 네 개 정도일 경우가 가격 대 성능비 면에서 가장 유리한 것으로 연구되었다.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.12 no.4
• /
• pp.1889-1896
• /
• 2011

There has been a lot of researches to develop techniques that provide redundant paths, there by making Multistage Interconnection Networks(MINs) fault tolerant. So far, the redundant paths in MINs have been realized by adding additional hardware such as extra stages or duplicated data links. This paper presents a new MIN topology called Hierarchical MIN. The proposed MIN is constructed with 2.5N-4 switching elements, which are much fewer than that of the classical MINs. Even though there are fewer hardware than the classical MINs, the HMIN possesses the property of full access and also provides alternative paths for the fault tolerant. Furthermore, since there is the short cut in HMIN for the localized communication, it takes advantage of exploiting the locality of reference in multiprocessor systems. Its performance under varying degrees of localized communication is analysed and simulated.

• Proceedings of the Korean Information Science Society Conference
• /
• 2005.11a
• /
• pp.166-168
• /
• 2005

본 논문에서는, axa 출력 버퍼 스위치로 구성된 핫스팟이 발생된 상황 하에서 다단 연결 망(Multistage interconnection Network, MIN)의 성능 예측 모형을 제안하였다. 제안한 성능 예측 오형은 먼저 네트워크 내부 임의 스위치 입력 단에 유입되는 데이터 패킷이 스위치 내부에서 전송되는 유형을 확률적으로 분석하여 설계하였다. 성능분석 모형은 스위치에 장착된 버퍼의 개수와 무관하게 버퍼를 장착한 axa 스위치의 성능, 네트워크 정상상태 처리율(Normalized Throughput, NT)과 네트워크 지연시간(Network Delay)의 예측이 가능하고, 나아가서 이들로 구성진 모든 종류의 다단 연결망 성능 분석에 적용이 용이하다. 제안한 수학적 성능 분석 연구의 실효성 검증을 위하여 병행된 시뮬레이션 결과는 상호 미세한 오차 범위 내에서 모형의 예측 데이터와 일치하는 결과를 보여 분석 모형의 타당성을 입증하였다.

• ETRI Journal
• /
• v.45 no.1
• /
• pp.60-74
• /
• 2023

Integration of the machine learning (ML) technique in all-optical networks can enhance the effectiveness of resource utilization, quality of service assurances, and scalability in optical networks. All-optical multistage interconnection networks (MINs) are implicitly designed to withstand the increasing highvolume traffic demands at data centers. However, the contention resolution mechanism in MINs becomes a bottleneck in handling such data traffic. In this paper, a select list of ML algorithms replaces the traditional electronic signal processing methods used to resolve contention in MIN. The suitability of these algorithms in improving the performance of the entire network is assessed in terms of injection rate, average latency, and latency distribution. Our findings showed that the ML module is recommended for improving the performance of the network. The improved performance and traffic grooming capabilities of the module are also validated by using a hardware testbed.

• The Transactions of the Korea Information Processing Society
• /
• v.6 no.5
• /
• pp.1212-1218
• /
• 1999

This paper proposes a new multipath interconnection network(MIN) structure for improving performance which uses widely in the design of multiprocessing, ATM system and VOD server. For improving performance such as passthrough ratio and packet latency, it proposes new routing method which routes one of the collided packets into the i+1-th switching stage of the adding Banyan MIN network when it occurred collision at the i-th switching stage of the basic MIN again when they collide each other. The new improved performance MIN network has been compared with MBSF, TBSF and PBSF structured MIN network from the viewpoint of passthrough ratio and the number of switching stage vs. passthrough ratio. It is shown to improve a performance and to be a simple structure which reducing the number of switching stage of adding MIN in comparison with other structured MIN including TBSF.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.37 no.6
• /
• pp.1-9
• /
• 2000

This paper presents a Cluster Oriented Multistage Interconnection Network called COMR. COMR can be constructed suitable for the parallel application with localized communication by providing the shortcut path inside the processor-memory cluster which has frequent data communication. We evaluate the performance of COMR with respect to probability of acceptance, bandwidth, cost-effectiveness and average distance under varying degrees of localized communication. According to the result of analysis for performance evaluation, COMR shows higher performance than the regular MINs of the same network size in the highly localized communication. In the worst case, the diameter of an N$\times$N COMR is only n+1 which has only one stage more as compared the MIN with the same network size. Therefore COMR can be used as an attractive interconnection network for parallel applications with not only the localized communication distribution but also the uniform distribution in shared-memory multiprocessor system.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.21 no.2
• /
• pp.141-170
• /
• 1996

The purpose of this paper is to critically review and compare the evaluation methods of fault tolerant MIN system and to suggest future study direction. The reliability analysis of MIN system is discussed. The system performance modeling under the faulty components is another main focus of this study.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2002.04a
• /
• pp.349-352
• /
• 2002

다단계 상호연결 네트워크 (multistage interconnection network : MIN) 동작 연구에 있어서 가장 중요한 것 중의 하나가 네트워크 사이클의 운영 방법이라 할 수 있다. 일반적인 MIN 기반 교환기 구조에서 최악의 경우는 수신 버퍼 모듈이 비어있어서 셀이 가장 앞쪽의 버퍼 위치로 움직여야만 하는 상황에서 발생한다. 그러나 실제적으로는 입력 로드의 실질적인 크기에 대해서 대부분의 버퍼 모듈들은 거의 항상 확 차게 된다. 그렇기 때문에 일반적인 체계의 긴 네트워크 사이클은 대역폭의 낭비를 가져오게 된다. 본 논문에서는 이런 문제점을 수정하여 고안한 단위 단계 버퍼링 체계를 사용하여 입력 버퍼에 관한 분석 모델을 제안하였다. 이 제안된 체계는 매개변수의 수와 상태전이 다이어그램에 있어서 기존의 모델링 기법들보다 매우 단순하다. 제안된 모델과 시뮬레이션의 처리율 및 지연시간의 비교를 통해서 제안된 모델의 정확성을 검증하였다.

• Journal of Korea Multimedia Society
• /
• v.2 no.2
• /
• pp.184-194
• /
• 1999

We present a new algorithm to minimize channel contention while sending multiple messages from multiple source to overlapped destination set on Multistage Interconnection Network (MIN) which supports wormhole routed turnaround routing. The multicast tree of the U-MIN(Unicast MIN) algorithm is useful in performing messages from one source to multiple destination but gives rise to a serious channel connection in performing multiple multicast because it has been designed for only single multicast. For multiple multicast communication on MIN, we address how to implement multiple multicast services efficiently. And a SPU-MIN(Source Partitioned Unicast MIN) algorithm is proposed and shown to be superior than the U-MIN algorithm for multiple multicast. The turnaround routing algorithm based on wormhole routing technique is employed as a message sending method.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.28B no.10
• /
• pp.807-821
• /
• 1991

A mulitpath MIN(Multistage Interconnection Network), CSMP(Chained Shuffle Multi-Path) network, is proposed, having fault-tolerance and dynamic reroutability. The number of stages and the number of links between adjacent stagges are the same as in single path MINs, so the overall hardware complexity is considerably reduced in comparison with other multipath MINs. The CSMP networks feature links between switches belonging to the same state, forming loops of switches. The network can tolerate multiple faults, up to (N/4)*(log$_2$N-1), having occured in any stages including the first and the last ones(N:NO. of input). To analyze reliability, terminal reliability (TR) and mean time to failure( MTTE) age given for the networks, and the TR figures are compared to those of other static and dynamic rerouting multipath MINs. Also the MTTE figures are compared. The performance of the proposed network with respect to its bandwidth (BW) and probability of acceptance(PA) is analyzed and is compared to that of other more complex multipath MINs. The cost efficiency analysis of reliability and performance shows that the network is more cost-effective than other previously proposed fault-tolerant multipath MINs.