Ma Xiang-jie, Li Xiu-qin, Lan Ju-long, Zhang Bai-sheng. Graphic Model and Performance Analysis of a Novel Scalable Multiple-Plane and Multiple-Stage Packet Switching Fabric[J]. Journal of Electronics & Information Technology, 2009, 31(5): 1026-1030. doi: 10.3724/SP.J.1146.2008.00483
Citation:
Ma Xiang-jie, Li Xiu-qin, Lan Ju-long, Zhang Bai-sheng. Graphic Model and Performance Analysis of a Novel Scalable Multiple-Plane and Multiple-Stage Packet Switching Fabric[J]. Journal of Electronics & Information Technology, 2009, 31(5): 1026-1030. doi: 10.3724/SP.J.1146.2008.00483
Ma Xiang-jie, Li Xiu-qin, Lan Ju-long, Zhang Bai-sheng. Graphic Model and Performance Analysis of a Novel Scalable Multiple-Plane and Multiple-Stage Packet Switching Fabric[J]. Journal of Electronics & Information Technology, 2009, 31(5): 1026-1030. doi: 10.3724/SP.J.1146.2008.00483
Citation:
Ma Xiang-jie, Li Xiu-qin, Lan Ju-long, Zhang Bai-sheng. Graphic Model and Performance Analysis of a Novel Scalable Multiple-Plane and Multiple-Stage Packet Switching Fabric[J]. Journal of Electronics & Information Technology, 2009, 31(5): 1026-1030. doi: 10.3724/SP.J.1146.2008.00483
A novel scalable Multiple-Plane and Multiple-Stage (MPMS) packet switching fabric is proposed in this paper. Firstly, the graphic model of the MPMS fabric is built. The neighbouring connectivity and reachability of MPMS are described. Vertexes are classified into balanced vertex and competitive vertex, and the non-blocking condition is proved, which may determine its switching performance. The MPMS fabric is proved to achieve a P times maximum port rate, a square of maximum port count, only linearly increasing structure complexity, and therefore it is much more scalable.
Wang F, Zhu Wen-qi, and Hamdi M. The Central-stagebuffered Clos-network to emulate an OQ switch[C]. IEEEGlobecom Proceedings, California, USA, Nov. 2006:4244-4257.[2]McKeown N. The iSLIP scheduling algorithm forinput-queued switches[J].IEEE/ACM Trans. on Networking.1999, 7(2):188-200[3]Mekkittikul A and McKeown N. A practical schedulingalgorithm for achieving 100% throughput in input-queuedswitches[C]. Proceeding of IEEE Infocom, San Francisco,USA, 2004: 792-799.[4]Clos C. A study of non-blocking switching networks[J]. BellSystems Technical Journal, 1953, 32(1): 406-424.[5]Goke L R and Lipovski G J. Banyan networks forpartitioning processor systems[C]. Proceeding Annual SympComputer Architecture, USA, Dec. 2003: 21-28.[6]Wu C L and Feng T Y. On a class of multistageinterconnection networks[J].IEEE Trans. on Computers.2000, 29(8):694-702[7]Pease M C. The indirect binary n-Cube microprocessorarray[J].IEEE Trans. on Computers.2005, 26(5):458-473[8]Iyer S and McKeown N. Making parallel packet switchespractical[C]. Proceeding of IEEE Infocom, USA, 2001, Vol.3:1680-1687.