About | Contact Us | Register | Login
ProceedingsSeriesJournalsSearchEAI
Proceedings of the Workshop on ns-3

Research Article

A SACK-based Conservative Loss Recovery Algorithm for ns-3 TCP: a Linux-inspired Proposal

Cite
BibTeX Plain Text
  • @INPROCEEDINGS{10.1145/3067665.3067666,
        author={Natale Patriciello},
        title={A SACK-based Conservative Loss Recovery Algorithm for ns-3 TCP: a Linux-inspired Proposal},
        proceedings={Proceedings of the Workshop on ns-3},
        publisher={ACM},
        proceedings_a={WNS3},
        year={2017},
        month={7},
        keywords={TCP SACK ns-3},
        doi={10.1145/3067665.3067666}
    }
    
  • Natale Patriciello
    Year: 2017
    A SACK-based Conservative Loss Recovery Algorithm for ns-3 TCP: a Linux-inspired Proposal
    WNS3
    ACM
    DOI: 10.1145/3067665.3067666
Natale Patriciello1,2,*
  • 1: Department of Engineering Enzo Ferrari, University of Modena and Reggio Emilia
  • 2: Via P. Vivarelli 10 Modena, Italy 41125
*Contact email: natale.patriciello@unimore.it

Abstract

This paper presents a new implementation of the SACK option and RFC 6675 loss recovery algorithm in ns-3. As is already happening in the Linux kernel, we have merged the loss recovery algorithm in the part of the code shared by all TCP variants. This merge allows bringing the performance improvements under correlated losses to all congestion control algorithms. The paper then shows the perfor-mance of TCP Hybla, Highspeed, and Vegas, under a scenario with different packet error rates, comparing the performance obtained with and without the use of the TCP SACK option. The results allow concluding that the main characteristics of the congestion control algorithms are untouched, but thanks to the implemented option their resilience to correlated losses is improved.

Keywords
TCP SACK ns-3
Published
2017-07-27
Publisher
ACM
http://dx.doi.org/10.1145/3067665.3067666
Copyright © 2017–2025 ACM
EBSCOProQuestDBLPDOAJPortico
EAI Logo

About EAI

  • Who We Are
  • Leadership
  • Research Areas
  • Partners
  • Media Center

Community

  • Membership
  • Conference
  • Recognition
  • Sponsor Us

Publish with EAI

  • Publishing
  • Journals
  • Proceedings
  • Books
  • EUDL