EXIO : Enabling globus on rdma networks-a case study with InfiniBand

For its low-latency, high bandwidth, and low CPU utilization, RDMA (Remote Direct Memory Access) has established itself as an effective data movement technology in many networking environments. This includes recent incarnations of InfiniBand-based RDMA on longdistance networks such as UltraScience Net. However, grid runt-time systems, such as the Globus toolkit, are not yet capable of utilizing RDMA for fast data movement. In this paper, we examine the architecture of Globus for the feasibility of enabling RDMA. An I/O architecture EXIO is designed to extend the Globus XIO system and match the characteristics of RDMA in terms of both connection establishment and communication progress. An initial proof-of-concept implementation of EXIO is realized on InfiniBand. Our experimental results demonstrate that, compared to IPoIB and 10GigE, RDMA can significantly improve the performance of Globus, reducing the latency by 32% and increasing the bandwidth by more than three times. In achieving such performance improvements, RDMA dramatically cuts down CPU utilization of Globus clients and servers. These results adequately demonstrate that EXIO is effectively designed and implemented to exploit the benefits of RDMA for Globus.

[1]  Dhabaleswar K. Panda,et al.  PVFS over InfiniBand: design and performance evaluation , 2003, 2003 International Conference on Parallel Processing, 2003. Proceedings..

[2]  Robert B. Ross,et al.  PVFS: A Parallel File System for Linux Clusters , 2000, Annual Linux Showcase & Conference.

[3]  Hiroyuki Ohsaki,et al.  GridFTP-APT: automatic parallelism tuning mechanism for data transfer protocol GridFTP , 2006, Sixth IEEE International Symposium on Cluster Computing and the Grid (CCGRID'06).

[4]  George Almási,et al.  Scalable RDMA performance in PGAS languages , 2009, 2009 IEEE International Symposium on Parallel & Distributed Processing.

[5]  Jeffrey S. Vetter,et al.  Experimental Analysis of InfiniBand Transport Services on WAN , 2008, 2008 International Conference on Networking, Architecture, and Storage.

[6]  Stephen Bailey,et al.  An Overview of RDMA over IP , 2002 .

[7]  Jeffrey S. Vetter,et al.  Wide-area performance profiling of 10GigE and InfiniBand technologies , 2008, 2008 SC - International Conference for High Performance Computing, Networking, Storage and Analysis.

[8]  Peter F. Corbett,et al.  The Direct Access File System , 2003, FAST.

[9]  Renato Recio,et al.  An RDMA Protocol Specification , 2002 .

[10]  Dhabaleswar K. Panda,et al.  High performance support of parallel virtual file system (PVFS2) over Quadrics , 2005, ICS '05.

[11]  Nageswara S. V. Rao,et al.  Experimental evaluation of infiniband transport over local- and wide-area networks , 2007, SpringSim '07.

[12]  Brent Callaghan,et al.  NFS over RDMA , 2003, NICELI '03.

[13]  Amith R. Mamidala,et al.  Designing Efficient FTP Mechanisms for High Performance Data-Transfer over InfiniBand , 2009, 2009 International Conference on Parallel Processing.

[14]  William E. Allcock,et al.  The globus extensible input/output system (XIO): a protocol independent IO system for the grid , 2005, 19th IEEE International Parallel and Distributed Processing Symposium.

[15]  D. Martin Swany,et al.  Improving GridFTP performance using the Phoebus session layer , 2009, Proceedings of the Conference on High Performance Computing Networking, Storage and Analysis.

[16]  Chase Qishi Wu,et al.  Ultrascience net: network testbed for large-scale science applications , 2005, IEEE Communications Magazine.

[17]  Chase Qishi Wu,et al.  UltraScience Net: High-Performance Network Research Test-Bed , 2009, Int. J. Distributed Sens. Networks.

[18]  William E. Allcock,et al.  The Globus Striped GridFTP Framework and Server , 2005, ACM/IEEE SC 2005 Conference (SC'05).

[19]  Kai Li,et al.  Experiences with VI communication for database storage , 2002, Proceedings 29th Annual International Symposium on Computer Architecture.

[20]  Hemal Shah,et al.  A study of iSCSI extensions for RDMA (iSER) , 2003, NICELI '03.

[21]  R. Kettimuthu UDT as an Alternative Transport Protocol for GridFTP , 2009 .

[22]  Robert L. Grossman,et al.  Udt: a high performance data transport protocol , 2005 .

[23]  P. Wyckoff,et al.  iSER Storage Target for Object-Based Storage Devices , 2007, Fourth International Workshop on Storage Network Architecture and Parallel I/Os (SNAPI 2007).

[24]  Dhabaleswar K. Panda,et al.  Performance of HPC Middleware over InfiniBand WAN , 2008, 2008 37th International Conference on Parallel Processing.

[25]  Dhabaleswar K. Panda,et al.  High performance RDMA-based MPI implementation over InfiniBand , 2003, ICS.