Design and testbed evaluation of RDMA-based middleware for high-performance data transfer applications

被引:6
作者
Ren, Yufei [1 ]
Li, Tan [1 ]
Yu, Dantong [2 ,3 ]
Jin, Shudong [1 ,4 ]
Robertazzi, Thomas [1 ,5 ]
机构
[1] SUNY Stony Brook, Stony Brook, NY 11794 USA
[2] Brookhaven Natl Lab, Upton, NY 11973 USA
[3] Brookhaven Natl Lab, Grid Comp Grp, Upton, NY 11973 USA
[4] SUNY Stony Brook, Ctr Wireless & Informat Technol, Stony Brook, NY 11794 USA
[5] SUNY Stony Brook, Dept Elect & Comp Engn, Stony Brook, NY 11794 USA
基金
美国能源部;
关键词
Distributed systems; Middleware; Remote Direct Memory Access;
D O I
10.1016/j.jss.2013.01.070
中图分类号
TP31 [计算机软件];
学科分类号
081202 ; 0835 ;
摘要
Providing high-speed data transfer is vital to various data-intensive applications supported by data center networks. We design a middleware layer of high-speed communication based on Remote Direct Memory Access (RDMA) that serves as the common substrate to accelerate various data transfer tools, such as FTP, HTTP, file copy, sync and remote file I/O. This middleware offers better end-to-end bandwidth performance than the traditional TCP-based alternatives, while it hides the heterogeneity of the underlying high-speed architecture. This paper describes this middleware's function modules, including resource abstraction and task synchronization and scheduling, that maximize the parallelism and performance of RDMA operations. For networks without RDMA hardware acceleration, we integrate Linux kernel optimization techniques to reduce data copy and processing in the middleware. We provide a reference implementation of the popular file-transfer protocol over this RDMA-based middleware layer, called RFTP. Our experimental results show that our RFTP outperforms several TCP-based FTP tools, such as GridFTP, while it maintains very low CPU consumption on a variety of data center platforms. Furthermore, those results confirm that our RFTP tool achieves near line-speed performance in both LAN and WAN, and scales consistently from 10 Gbps Ethernet to 40 Gbps Ethernet and InfiniBand environments. (C) 2013 Elsevier Inc. All rights reserved.
引用
收藏
页码:1850 / 1863
页数:14
相关论文
共 28 条
  • [1] [Anonymous], 2011, LUSTR FIL SYST
  • [2] Bierbaum N, 2002, CONF LOCAL COMPUT NE, P733, DOI 10.1109/LCN.2002.1181853
  • [3] Remote Direct Memory Access over the Converged Enhanced Ethernet Fabric: Evaluating the Options
    Cohen, David
    Talpey, Thomas
    Kanevsky, Arkady
    Cummings, Uri
    Krause, Michael
    Recio, Renato
    Crupnicoff, Diego
    Dickman, Lloyd
    Grun, Paul
    [J]. 2009 17TH IEEE SYMPOSIUM ON HIGH-PERFORMANCE INTERCONNECTS (HOTI 2009), 2009, : 123 - +
  • [4] Danalis A., 2008, P IEEE INT S PAR DIS
  • [5] Danalis A., 2008, EUROPVM MPI 2008
  • [6] *DAT COLL, 2002, UDAPL US DIR ACC PRO
  • [7] ESnet, 2012, LIN TCP TUN
  • [8] Feng W., 2005, P 13 S HIGH PERF INT
  • [9] Frey P. W., 2009, P IEEE INT C DISTR C
  • [10] Globus, 2012, GT 4 0 GRIDFTP GLOSS