Eager Data Transfer Mechanism for Reducing Communication Latency in User-Level Network Protocols

被引:2
|
作者
Won, Chulho [1 ]
Lee, Ben [2 ]
Park, Kyoung [3 ]
Kim, Myung-Joon [3 ]
机构
[1] Calif State Univ, Elect & Comp Engn Dept, Fresno, CA USA
[2] Oregon State Univ, Corvallis, OR 97331 USA
[3] Elect & Telecommun Res Inst, Daejeon, South Korea
来源
关键词
Data Transfer; Cache Coherence; User-Level; Low-Latency; Network Protocols; Message; VIA;
D O I
10.3745/JIPS.2008.4.4.133
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Clusters have become a popular alternative for building high-performance parallel computing systems. Today's high-performance system area network (SAN) protocols such as VIA and IBA significantly reduce user-to-user communication latency by implementing protocol stacks outside of operating system kernel. However, emerging parallel applications require a significant improvement in communication latency. Since the time required for transferring data between host memory and network interface (NI) make up a large portion of overall communication latency, the reduction of data transfer time is crucial for achieving low-latency communication. In this paper, Eager Data Transfer (EDT) mechanism is proposed to reduce the time for data transfers between the host and network interface. The EDT employs cache coherence interface hardware to directly transfer data between the host and NI. An EDT-based network interface was modeled and simulated on the Linux-based, complete system simulation environment, Linux/SimOS. Our simulation results show that the EDT approach significantly reduces the data transfer time compared to DMA-based approaches. The EDT-based NI attains 17% to 38% reduction in user-to-user message time compared to the cache-coherent DMA-based NIs for a range of message sizes (64 bytes similar to 4 Kbytes) in a SAN environment.
引用
收藏
页码:133 / 144
页数:12
相关论文
共 50 条
  • [1] User-level network interface protocols
    Bhoedjang, RAF
    Ruhl, T
    Bal, HE
    COMPUTER, 1998, 31 (11) : 53 - +
  • [2] Reducing waiting costs in user-level communication
    Damianakis, SN
    Chen, YQ
    Felten, EW
    11TH INTERNATIONAL PARALLEL PROCESSING SYMPOSIUM, PROCEEDINGS, 1997, : 381 - 387
  • [3] ATM and Fast Ethernet network interfaces for user-level communication
    Welsh, M
    Basu, A
    vonEicken, T
    THIRD INTERNATIONAL SYMPOSIUM ON HIGH-PERFORMANCE COMPUTER ARCHITECTURE - PROCEEDINGS, 1997, : 332 - 342
  • [4] Signals, timers, and continuations for multithreaded user-level protocols
    Gomez, JC
    Ramos, JR
    Rego, V
    SOFTWARE-PRACTICE & EXPERIENCE, 2006, 36 (05): : 449 - 471
  • [6] Gang scheduling with lightweight user-level communication
    Frachtenberg, E
    Petrini, F
    Coll, S
    Feng, WC
    INTERNATIONAL CONFERENCE ON PARALLEL PROCESSING WORKSHOPS, PROCEEDINGS, 2001, : 339 - 345
  • [7] User-level Data Center Tomography
    Twigg, Neil Alexander
    Fayed, Marwan
    Perkins, Colin
    Pezaros, Dimitrios
    Tso, Posco
    ACM SIGCOMM COMPUTER COMMUNICATION REVIEW, 2012, 42 (04) : 101 - 102
  • [8] A user-level approach to network attached storage
    Sohan, R
    Hand, S
    LCN 2005: 30TH CONFERENCE ON LOCAL COMPUTER NETWORKS, PROCEEDINGS, 2005, : 108 - 114
  • [9] Memory management for user-level network interfaces
    Welsh, M
    Basu, A
    Huang, XW
    von Eicken, T
    IEEE MICRO, 1998, 18 (02) : 77 - 82
  • [10] USER-LEVEL INTERPROCESS COMMUNICATION FOR SHARED MEMORY MULTIPROCESSORS
    BERSHAD, BN
    ANDERSON, TE
    LAZOWSKA, ED
    LEVY, HM
    ACM TRANSACTIONS ON COMPUTER SYSTEMS, 1991, 9 (02): : 175 - 198