Journal of Computer Science and Technology
Quick Search in JCST
 Advanced Search 
      Home | PrePrint | SiteMap | Contact Us | FAQ
 
Indexed by   SCIE, EI ...
Bimonthly    Since 1986
Journal of Computer Science and Technology 2014, Vol. 29 Issue (2) :281-292    DOI: 10.1007/s11390-014-1430-0
Special Section on Cloud-Sea Computing Systems Current Issue | Archive | Adv Search << Previous Articles | Next Articles >>
A High-Performance and Cost-Effcient Interconnection Network for High-Density Servers
Wen-Tao Bao1, 2 (包雯韬), Student Member, CCF, ACM, IEEE Bin-Zhang Fu1 (付斌章), Member, CCF, ACM, IEEE, Ming-Yu Chen1, 2 (陈明宇), Member, CCF, ACM, IEEE and Li-Xin Zhang1, 2 (张立新), Member, ACM, IEEE
1 State Key Laboratory of Computer Architecture, Institute of Computing Technology, Chinese Academy of Sciences Beijing 100190, China;
2 University of Chinese Academy of Sciences, Beijing 100049, China

Abstract
Reference
Related Articles
Download: [PDF 1784KB]     Export: BibTeX or EndNote (RIS)  
Abstract The high-density server is featured as low power, low volume, and high computational density. With the rising use of high-density servers in data-intensive and large-scale web applications, it requires a high-performance and cost-effcient intra-server interconnection network. Most of state-of-the-art high-density servers adopt the fully-connected intra-server network to attain high network performance. Unfortunately, this solution costs too much due to the high degree of nodes. In this paper, we exploit the theoretically optimized Moore graph to interconnect the chips within a server. Accounting for the suitable size of applications, a 50-size Moore graph, called Hoffman-Singleton graph, is adopted. In practice, multiple chips should be integrated onto one processor board, which means that the original graph should be partitioned into homogeneous connected subgraphs. However, the existing partition scheme does not consider above problem and thus generates heterogeneous subgraphs. To address this problem, we propose two equivalent-partition schemes for the Hoffman-Singleton graph. In addition, a logic-based and minimal routing mechanism, which is both time and area effcient, is proposed. Finally, we compare the proposed network architecture with its counterparts, namely the fully-connected, Kautz and Torus networks. The results show that our proposed network can achieve competitive performance as fully-connected network and cost close to Torus.
Articles by authors
Wen-Tao Bao
Bin-Zhang Fu
Ming-Yu Chen
Li-Xin Zhang
Keywordshigh-density server   interconnection network   Moore graph   Hoffman-Singleton graph   equivalent partition     
Received 2013-11-14;
Fund:

This work was supported by the Strategic Priority Research Program of the Chinese Academy of Sciences under Grant No. XDA06010401, the National Natural Science Foundation of China under Grant Nos. 61202056, 61331008, 61221062, and the Huawei Research Program of China under Grant No. YBCB2011030.

About author: Wen-Tao Bao received the B.S. degree from Jilin University, Changchun, in 2012. Now she is pursuing her M.S. degree in Institute of Computing Technology, Chinese Academy of Sciences, Beijing. Her research interests include highperformance and high-reliable interconnection networks.
Cite this article:   
Wen-Tao Bao, Bin-Zhang Fu, Ming-Yu Chen, Li-Xin Zhang.A High-Performance and Cost-Effcient Interconnection Network for High-Density Servers[J]  Journal of Computer Science and Technology, 2014,V29(2): 281-292
URL:  
http://jcst.ict.ac.cn:8080/jcst/EN/10.1007/s11390-014-1430-0
Copyright 2010 by Journal of Computer Science and Technology