• DocumentCode
    1778261
  • Title

    Demonstration of an ultra-low latency energy efficient internet for cloud computing

  • Author

    Szymanski, T.H.

  • Author_Institution
    THS Technol., McMaster Univ., Hamilton, ON, Canada
  • fYear
    2014
  • fDate
    10-13 Jan. 2014
  • Firstpage
    1122
  • Lastpage
    1123
  • Abstract
    An Enhanced-Internet network which supports ´ultra-low-latency´ energy efficient communications for cloud services is demonstrated. The Enhanced-Internet combines several technologies: (i) two non-interfering traffic classes, the Guaranteed-Rate (GR) and Best-Effort (BE) classes, (ii) a flowbased resource-reservation signalling protocol for the GR class, (iii) a QoS-aware low-jitter router scheduling algorithm with 100% bandwidth-efficiency, and (iv) aggregation and smoothening of cloud traffic, to achieve exceptionally low latencies and high energy efficiencies. The Enhanced-Internet supports ultralow latency machine-to-machine communications, for automated financial trading systems. Current cloud data-centers and the Internet infrastructure operate at low efficiencies, resulting in excess energy costs of several $US Billion per year. The Enhanced- Internet improves the efficiency of the current infrastructure, thereby reducing energy costs. A Big Data/Exascale Cloud Computing system using the Enhanced-Internet can achieve significantly improved performance and energy-efficiency, by utilizing the idle-periods in data-centers. The demonstration will illustrate a saturated Enhanced-Internet network supporting one thousand traffic flows, each receiving essentially-perfect service, with very low latencies and very high energy-efficiencies.
  • Keywords
    Big Data; Web services; aggregation; cloud computing; computer centres; power aware computing; protocols; quality of service; resource allocation; scheduling; telecommunication power management; telecommunication traffic; GR class; Internet infrastructure; QoS-aware low-jitter router scheduling algorithm; automated financial trading systems; bandwidth-efficiency; big data; cloud data-centers; cloud services; cloud traffic smoothening; enhanced-Internet network; exascale cloud computing system; flow-based resource-reservation signalling protocol; noninterfering traffic classes; ultralow latency energy efficient Internet; ultralow latency machine-to-machine communications; ultralow-latency energy efficient communications; Cloud computing; Delays; Quality of service; Routing; Routing protocols; QoS; big data; buffer sizes; cloud computing; data centers; energy efficiency; exascale; future internet; low jitter; low latency; machine-to-machine; openflow; routing; scheduling; software defined networks;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Consumer Communications and Networking Conference (CCNC), 2014 IEEE 11th
  • Conference_Location
    Las Vegas, NV
  • Print_ISBN
    978-1-4799-2356-4
  • Type

    conf

  • DOI
    10.1109/CCNC.2014.6994433
  • Filename
    6994433