Filters








114 Hits in 5.0 sec

Zero Copy Sockets Direct Protocol over InfiniBand — Preliminary Implementation and Performance Analysis

D. Goldenberg, M. Kagan, R. Ravid, M.S. Tsirkin
13th Symposium on High Performance Interconnects (HOTI'05)  
Sockets Direct Protocol (SDP) is a byte-stream transport protocol implementing the TCP SOCK_STREAM semantics utilizing transport offloading capabilities of the InfiniBand fabric.  ...  This paper presents the major architectural aspects of the SDP protocol, the ZCopy implementation, and a preliminary performance evaluation.  ...  The most commonly used protocols are Message Passing Interface (MPI), IP over InfiniBand (IPoIB), Sockets Direct Protocol (SDP), SCSI RDMA Protocol (SRP), and Direct Access Provider Library (DAPL).  ... 
doi:10.1109/conect.2005.35 dblp:conf/hoti/GoldenbergKRT05 fatcat:6rxyjjlqbzhgvb5a4ndi775vvy

Architecture and Implementation of Sockets Direct Protocol in Windows

Dror Goldenberg, Tzachi Dar, Gilad Shainer
2006 2006 IEEE International Conference on Cluster Computing  
Sockets Direct Protocol (SDP) is a byte stream protocol that utilizes the capabilities of the InfiniBand fabric to transparently achieve performance gains for existing socket-based networked applications  ...  We go through the motivation, the main implementation architectural aspects and challenges. We present preliminary performance results.  ...  Previous SDP research [6] [7] [9] [19] mainly focused on development of SDP for the Linux operating system, performance analysis, tuning and addition of Zero Copy path to the SDP implementations.  ... 
doi:10.1109/clustr.2006.311918 dblp:conf/cluster/GoldenbergDS06 fatcat:lxeazuq7dvfftflurmdkc3scui

RDMA over Ethernet — A preliminary study

Hari Subramoni, Ping Lai, Miao Luo, Dhabaleswar K. Panda
2009 2009 IEEE International Conference on Cluster Computing and Workshops  
In this paper, we present a detailed evaluation and analysis of the new RDMAoE protocol as opposed to the earlier overlay protocols as well as native-IB and socket based implementations.  ...  high performance capabilities that InfiniBand offers.  ...  Other than the efforts that try to use zero-copy send/receive and RDMA technologies over Ethernet, there are also other works that have been done on different protocols.  ... 
doi:10.1109/clustr.2009.5289144 dblp:conf/cluster/SubramoniLLP09 fatcat:ukfy55u3lbew3bwzv3m2fdlwae

Java Fast Sockets: Enabling high-speed Java communications on high performance clusters

Guillermo L. Taboada, Juan Touriño, Ramón Doallo
2008 Computer Communications  
The Socket Direct Protocol (SDP) over Infiniband [11] is the representative socket library of the Offload Sockets Framework (OSF).  ...  A direct implementation of native sockets on top of low-level communication libraries can avoid the TCP/IP overhead, and thus increases performance. Representative examples are next presented.  ...  Acknowledgement This work was funded by the Ministry of Education and Science of Spain under Project TIN2004-07797-C02 and by the Galician Government (Xunta de Galicia) under Project PGIDIT06PXIB 105228PR  ... 
doi:10.1016/j.comcom.2008.08.012 fatcat:jtribpsapzdklcw2fkmvnto3qy

10-Gigabit iWARP Ethernet: Comparative Performance Analysis with InfiniBand and Myrinet-10G

Mohammad J. Rashti, Ahmad Afsahi
2007 2007 IEEE International Parallel and Distributed Processing Symposium  
At the MPI level, iWARP performs better than InfiniBand in queue usage and buffer re-use.  ...  existing Ethernet infrastructure and protocols.  ...  We intend to extend our study to include uDAPL, sockets, and applications. We would also like to enhance the NetEffect MPI implementation.  ... 
doi:10.1109/ipdps.2007.370480 dblp:conf/ipps/RashtiA07 fatcat:fznac5yc3bchvn5on7tppbzofa

Protocols for wide-area data-intensive applications: Design and performance issues

Yufei Ren, Tan Li, Dantong Yu, Shudong Jin, Thomas Robertazzi, Brian L. Tierney, Eric Pouyoul
2012 2012 International Conference for High Performance Computing, Networking, Storage and Analysis  
We demonstrate its performance benefit on various local and wide-area testbeds, including the DOE ANI testbed with RoCE links and InfiniBand links.  ...  In this paper, we address the design choices to develop such protocols. We describe a protocol implemented as part of a communication middleware.  ...  OFED software also offers several middleware packages, such as IP over IB [10] and Sockets Direct Protocol (SDP) [11] , to allow socket-based applications to run over RDMA devices without needed to  ... 
doi:10.1109/sc.2012.98 dblp:conf/sc/RenLYJRTP12 fatcat:qjdz7xo6prckvepjlscvbcyn3q

A comprehensive zero-copy architecture for high performance distributed data acquisition over advanced network technologies for the CMS experiment

Gerry Bauer, Ulf Behrens, James Branson, Sebastian Bukowiec, Olivier Chaze, Sergio Cittolin, Jose Antonio Coarasa, Christian Deldicque, Marc Dobson, Aymeric Dupont, Samim Erhan, Dominique Gigi (+23 others)
2012 2012 18th IEEE-NPSS Real Time Conference  
This paper outlines a software architecture where zero-copy operations are used comprehensively at every processing point from the Application layer to the Physical layer.  ...  protocol facilitates deployment of several network technologies like Gigabit Ethernet, Myrinet, Infiniband etc.  ...  ACKNOWLEDGMENT This work was supported in part by the DOE and NSF (USA) and the Marie Curie Program.  ... 
doi:10.1109/rtc.2012.6418171 fatcat:wly4dhtolbanljgt5cweeiwayi

Designing next generation data-centers with advanced communication protocols and systems services

P. Balaji, K. Vaidyanathan, S. Narravula, H.-W. Jin, D.K. Panda
2006 Proceedings 20th IEEE International Parallel & Distributed Processing Symposium  
Current data-centers rely on TCP/IP over Fast-and Gigabit-Ethernet for data communication even within the cluster environment for costeffective designs, thus limiting their maximum capacity.  ...  We also present preliminary results in the various aspects of the framework, which demonstrate close to an order of magnitude performance benefits achievable by our framework as compared to existing data-centers  ...  BSDP and ZSDP are the copy-based SDP and synchronous zero-copy SDP implementations, respectively. We present ping-pong latency and uni-directional throughput micro-benchmarks results.  ... 
doi:10.1109/ipdps.2006.1639585 dblp:conf/ipps/BalajiVNJP06 fatcat:45vq7y2rdreqxdyh7u5w3hdqdq

F-MPJ: scalable Java message-passing communications on parallel systems

Guillermo L. Taboada, Juan Touriño, Ramón Doallo
2009 Journal of Supercomputing  
highperformance networks through the use of our high-performance Java sockets implementation (named JFS, Java Fast Sockets); (3) avoiding the use of communication buffers; and (4) optimizing MPJ collective  ...  A performance evaluation on an InfiniBand multi-core cluster has shown that F-MPJ communication primitives outperform representative MPJ libraries up to 60 times.  ...  Acknowledgements This work was funded by the Ministry of Education and Science of Spain under Projects TIN2004-07797-C02 and TIN2007-67537-C03-2 and by the Galician Government (Xunta de Galicia) under  ... 
doi:10.1007/s11227-009-0270-0 fatcat:2uddtgfvund3nhk2iuyimyouxi

Efficient wide area data transfer protocols for 100 Gbps networks and beyond

Ezra Kissel, Martin Swany, Brian Tierney, Eric Pouyoul
2013 Proceedings of the Third International Workshop on Network-Aware Data Management - NDM '13  
virtual circuit technology, and wide-area Remote Direct Memory Access (RDMA) protocols.  ...  In this paper we show that RDMA works well over wide-area virtual circuits, and uses much less CPU than TCP or UDP.  ...  A number of other RDMA and zero-copy protocols not involving InfiniBand (IB) have been proposed to run over Ethernet.  ... 
doi:10.1145/2534695.2534699 dblp:conf/sc/KisselSTP13 fatcat:tnos55l6l5dyjorpzswvbvmh7q

Energy-efficient data transfers in radio astronomy with software UDP RDMA

Przemyslaw Lenkiewicz, P. Chris Broekema, Bernard Metzler
2018 Future generations computer systems  
We have implemented a fully functional software prototype of such a protocol, supporting RDMA Read and Write operations and zero-copy capabilities.  ...  We present measurements of power consumption and achieved bandwidth and investigate the behaviour of all examined protocols when subjected to packet loss.  ...  Affairs (EL&I), and the Province of Drenthe.  ... 
doi:10.1016/j.future.2017.03.027 fatcat:gxrsheb6ufhtdmwetomx3kgwru

High Performance Data Transfer in Grid Environment Using GridFTP over InfiniBand

Hari Subramoni, Ping Lai, Raj Kettimuthu, Dhabaleswar K. Panda
2010 2010 10th IEEE/ACM International Conference on Cluster, Cloud and Grid Computing  
On the other hand, modern interconnects such as InfiniBand have introduced many advanced communication features, like zero-copy protocol and RDMA operations which can greatly improve the communication  ...  We also provide insights into the communication performance with some indepth analysis.  ...  In addition to providing high bandwidth and low latency, they also provide advanced features like zero-copy communication and Remote Direct Memory Access (RDMA) that enable the design of novel communication  ... 
doi:10.1109/ccgrid.2010.115 dblp:conf/ccgrid/SubramoniLKP10 fatcat:5ebdlsdzvzdtvaeatdnlza3iti

Analysis of the Effect of Core Affinity on High-Throughput Flows

Nathan Hanford, Vishal Ahuja, Matthew Farrens, Dipak Ghosal, Mehmet Balman, Eric Pouyoul, Brian Tierney
2014 2014 Fourth International Workshop on Network-Aware Data Management  
We conclude that affinitization has a significant impact on protocol processing efficiency, and that the performance bottleneck of the network receive process changes significantly with different affinitization  ...  We define protocol processing efficiency to be the amount of system resources (such as CPU and cache) used per unit of achieved throughput (in Gbps).  ...  Again, to ensure that the stress was placed on the end-system, the transfers were performed in zero-copy mode, which utilizes of the TCP sendfile system call to avoid unnecessary copies into and out of  ... 
doi:10.1109/ndm.2014.10 dblp:conf/sc/HanfordAFGBPT14 fatcat:v32oeknwp5fd7nuhtbkgwe64ya

Designing Efficient Systems Services and Primitives for Next-Generation Data-Centers

K. Vaidyanathan, S. Narravula, P. Balaji, D.K. Panda
2007 2007 IEEE International Parallel and Distributed Processing Symposium  
We present the performance results using data-center services such as cooperative caching and active resource monitoring and data-center primitives such as distributed data sharing substrate and distributed  ...  In this paper, we extend our previously proposed framework comprising of three layers (communication protocol support, data-center service primitives and advanced data-center services) that work together  ...  In [5] , we had demonstrated the potential of one such stack, namely the Sockets Direct Protocol (SDP), for data-center applications.  ... 
doi:10.1109/ipdps.2007.370507 dblp:conf/ipps/VaidyanathanNBP07 fatcat:xgvjpc3qjzh2fep5d6fukslupm

Towards In-transit Analysis on Supercomputing Environments [article]

Allan Santos and Hermano Lustosa and Fabio Porto and Bruno Schulze
2018 arXiv   pre-print
We use the Remote Direct Memory Access protocol (RDMA) that reduces the interference on performance caused by data copies and context switching.  ...  It adopts an in-memory data transfer strategy combined with TCP, using the BSD sockets API and the Linux splice(2) syscall. We present a performance evaluation with our work and traditional utilities.  ...  RDMA RDMA (Remote Direct Memory Access) is a communication protocol with more operations than the traditional BSD sockets API but using a more low-level approach.  ... 
arXiv:1805.06425v2 fatcat:u3rist3qajhldhbve5qbrao3ye
« Previous Showing results 1 — 15 out of 114 results