MURAL - Maynooth University Research Archive Library



    Approaching incast congestion with multi-host ethernet controllers


    Jereczek, Grzegorz and Lehmann-Miotto, Giovanna and Malone, David and Walukiewicz, Miroslaw (2017) Approaching incast congestion with multi-host ethernet controllers. In: 2017 IEEE Conference on Network Function Virtualization and Software Defined Networks (NFV-SDN), 6-8 November 2017, Berlin, Germany.

    [img]
    Preview
    Download (311kB) | Preview


    Share your research

    Twitter Facebook LinkedIn GooglePlus Email more...



    Add this article to your Mendeley library


    Abstract

    The bursty many-to-one communication pattern, typical for data acquisition systems, but also present in datacenter networks, is particularly demanding for commodity TCP/IP and Ethernet technologies. We expand our study of building incast-resistant networks based on software switches running on commercial-off-the-shelf servers. In this paper we provide the estimates for costs and physical area required to build such a network. Our estimates indicate that our proposed design offers significant cost advantage over traditional solutions, but higher space utilisation. Next, we show how the latter can be improved with multi-host Ethernet controllers, as an alternative to typical network interface cards. This can also make software switching easier to adapt in datacenter as a solution for incast congestion. We confirm the capabilities for incast-avoidance by evaluating the performance of a reference platform.

    Item Type: Conference or Workshop Item (Paper)
    Additional Information: This is the preprint version of the published paper, which is available at DOI: 10.1109/NFV-SDN.2017.8169824
    Keywords: computer centres; data acquisition; IP networks; local area networks; network interfaces; telecommunication congestion control; transport protocols;
    Academic Unit: Faculty of Science and Engineering > Mathematics and Statistics
    Faculty of Science and Engineering > Research Institutes > Hamilton Institute
    Item ID: 10058
    Depositing User: Dr. David Malone
    Date Deposited: 04 Oct 2018 14:20
    Refereed: Yes
    URI:

      Repository Staff Only(login required)

      View Item Item control page

      Downloads

      Downloads per month over past year