Heavy-hitter detection entirely in the data plane

Vibhaalakshmi Sivaraman, Srinivas Narayana, Ori Rottenstreich, Shanmugavelayutham Muthukrishnan, Jennifer Rexford

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Abstract

    Identifying the "heavy hitter" flows or flows with large traffic volumes in the data plane is important for several applications e.g., flow-size aware routing, DoS detection, and traffic engineering. However, measurement in the data plane is constrained by the need for linerate processing (at 10-100Gb/s) and limited memory in switching hardware. We propose HashPipe, a heavy hitter detection algorithm using emerging programmable data planes. HashPipe implements a pipeline of hash tables which retain counters for heavy flows while evicting lighter flows over time. We prototype HashPipe in P4 and evaluate it with packet traces from an ISP backbone link and a data center. On the ISP trace (which contains over 400,000 flows), we find that HashPipe identifies 95% of the 300 heaviest flows with less than 80KB of memory.

    Original languageEnglish (US)
    Title of host publicationSOSR 2017 - Proceedings of the 2017 Symposium on SDN Research
    PublisherAssociation for Computing Machinery, Inc
    Pages164-176
    Number of pages13
    ISBN (Electronic)9781450349475
    DOIs
    StatePublished - Apr 3 2017
    Event2017 Symposium on SDN Research, SOSR 2017 - Santa Clara, United States
    Duration: Apr 3 2017Apr 4 2017

    Publication series

    NameSOSR 2017 - Proceedings of the 2017 Symposium on SDN Research

    Conference

    Conference2017 Symposium on SDN Research, SOSR 2017
    CountryUnited States
    CitySanta Clara
    Period4/3/174/4/17

    Fingerprint

    Data storage equipment
    Computer hardware
    Pipelines
    Processing

    Keywords

    • Network algorithms
    • Network monitoring
    • Programmable networks
    • Software-defined networks

    ASJC Scopus subject areas

    • Computer Networks and Communications
    • Software

    Cite this

    Sivaraman, V., Narayana, S., Rottenstreich, O., Muthukrishnan, S., & Rexford, J. (2017). Heavy-hitter detection entirely in the data plane. In SOSR 2017 - Proceedings of the 2017 Symposium on SDN Research (pp. 164-176). (SOSR 2017 - Proceedings of the 2017 Symposium on SDN Research). Association for Computing Machinery, Inc. https://doi.org/10.1145/3050220.3063772

    Heavy-hitter detection entirely in the data plane. / Sivaraman, Vibhaalakshmi; Narayana, Srinivas; Rottenstreich, Ori; Muthukrishnan, Shanmugavelayutham; Rexford, Jennifer.

    SOSR 2017 - Proceedings of the 2017 Symposium on SDN Research. Association for Computing Machinery, Inc, 2017. p. 164-176 (SOSR 2017 - Proceedings of the 2017 Symposium on SDN Research).

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Sivaraman, V, Narayana, S, Rottenstreich, O, Muthukrishnan, S & Rexford, J 2017, Heavy-hitter detection entirely in the data plane. in SOSR 2017 - Proceedings of the 2017 Symposium on SDN Research. SOSR 2017 - Proceedings of the 2017 Symposium on SDN Research, Association for Computing Machinery, Inc, pp. 164-176, 2017 Symposium on SDN Research, SOSR 2017, Santa Clara, United States, 4/3/17. https://doi.org/10.1145/3050220.3063772
    Sivaraman V, Narayana S, Rottenstreich O, Muthukrishnan S, Rexford J. Heavy-hitter detection entirely in the data plane. In SOSR 2017 - Proceedings of the 2017 Symposium on SDN Research. Association for Computing Machinery, Inc. 2017. p. 164-176. (SOSR 2017 - Proceedings of the 2017 Symposium on SDN Research). https://doi.org/10.1145/3050220.3063772
    Sivaraman, Vibhaalakshmi ; Narayana, Srinivas ; Rottenstreich, Ori ; Muthukrishnan, Shanmugavelayutham ; Rexford, Jennifer. / Heavy-hitter detection entirely in the data plane. SOSR 2017 - Proceedings of the 2017 Symposium on SDN Research. Association for Computing Machinery, Inc, 2017. pp. 164-176 (SOSR 2017 - Proceedings of the 2017 Symposium on SDN Research).
    @inproceedings{3941d292de8f4c8db167151aa6bab848,
    title = "Heavy-hitter detection entirely in the data plane",
    abstract = "Identifying the {"}heavy hitter{"} flows or flows with large traffic volumes in the data plane is important for several applications e.g., flow-size aware routing, DoS detection, and traffic engineering. However, measurement in the data plane is constrained by the need for linerate processing (at 10-100Gb/s) and limited memory in switching hardware. We propose HashPipe, a heavy hitter detection algorithm using emerging programmable data planes. HashPipe implements a pipeline of hash tables which retain counters for heavy flows while evicting lighter flows over time. We prototype HashPipe in P4 and evaluate it with packet traces from an ISP backbone link and a data center. On the ISP trace (which contains over 400,000 flows), we find that HashPipe identifies 95{\%} of the 300 heaviest flows with less than 80KB of memory.",
    keywords = "Network algorithms, Network monitoring, Programmable networks, Software-defined networks",
    author = "Vibhaalakshmi Sivaraman and Srinivas Narayana and Ori Rottenstreich and Shanmugavelayutham Muthukrishnan and Jennifer Rexford",
    year = "2017",
    month = "4",
    day = "3",
    doi = "10.1145/3050220.3063772",
    language = "English (US)",
    series = "SOSR 2017 - Proceedings of the 2017 Symposium on SDN Research",
    publisher = "Association for Computing Machinery, Inc",
    pages = "164--176",
    booktitle = "SOSR 2017 - Proceedings of the 2017 Symposium on SDN Research",

    }

    TY - GEN

    T1 - Heavy-hitter detection entirely in the data plane

    AU - Sivaraman, Vibhaalakshmi

    AU - Narayana, Srinivas

    AU - Rottenstreich, Ori

    AU - Muthukrishnan, Shanmugavelayutham

    AU - Rexford, Jennifer

    PY - 2017/4/3

    Y1 - 2017/4/3

    N2 - Identifying the "heavy hitter" flows or flows with large traffic volumes in the data plane is important for several applications e.g., flow-size aware routing, DoS detection, and traffic engineering. However, measurement in the data plane is constrained by the need for linerate processing (at 10-100Gb/s) and limited memory in switching hardware. We propose HashPipe, a heavy hitter detection algorithm using emerging programmable data planes. HashPipe implements a pipeline of hash tables which retain counters for heavy flows while evicting lighter flows over time. We prototype HashPipe in P4 and evaluate it with packet traces from an ISP backbone link and a data center. On the ISP trace (which contains over 400,000 flows), we find that HashPipe identifies 95% of the 300 heaviest flows with less than 80KB of memory.

    AB - Identifying the "heavy hitter" flows or flows with large traffic volumes in the data plane is important for several applications e.g., flow-size aware routing, DoS detection, and traffic engineering. However, measurement in the data plane is constrained by the need for linerate processing (at 10-100Gb/s) and limited memory in switching hardware. We propose HashPipe, a heavy hitter detection algorithm using emerging programmable data planes. HashPipe implements a pipeline of hash tables which retain counters for heavy flows while evicting lighter flows over time. We prototype HashPipe in P4 and evaluate it with packet traces from an ISP backbone link and a data center. On the ISP trace (which contains over 400,000 flows), we find that HashPipe identifies 95% of the 300 heaviest flows with less than 80KB of memory.

    KW - Network algorithms

    KW - Network monitoring

    KW - Programmable networks

    KW - Software-defined networks

    UR - http://www.scopus.com/inward/record.url?scp=85018957154&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=85018957154&partnerID=8YFLogxK

    U2 - 10.1145/3050220.3063772

    DO - 10.1145/3050220.3063772

    M3 - Conference contribution

    T3 - SOSR 2017 - Proceedings of the 2017 Symposium on SDN Research

    SP - 164

    EP - 176

    BT - SOSR 2017 - Proceedings of the 2017 Symposium on SDN Research

    PB - Association for Computing Machinery, Inc

    ER -