TCP PLATO

Packet labelling to alleviate time-out

Shikhar Shukla, Shingau Chan, Adrian S W Tam, Abhishek Gupta, Yang Xu, H. Jonathan Chao

Research output: Contribution to journalArticle

Abstract

Many applications (e.g., cluster based storage and MapReduce) in modern data centers require a high fan-in, many-to-one type of data communication (known as TCP incast), which could cause severe incast congestion in switches and result in TCP goodput collapse, substantially degrading the application performance. The root cause of such a collapse is the long idle period of the Retransmission Timeout (RTO) that is triggered at one or more senders by packet losses in congested switches. In this paper we develop a packet labelling scheme PLATO, which improves the loss detection capabilities of NewReno using an innovative packet labelling system. Packets carrying this special label are preferentially enqueued, at the switch. This allows TCP to detect packet loss using three duplicate acknowledgements, instead of the time expensive RTO; thus avoiding the goodput collapse. PLATO makes minor modifications to NewReno and does not alter its congestion control mechanism. The implementation and simulations have been done in Network Simulator 3 (NS3). PLATO's performance is significantly better than NewReno as well as state-of-art incast solutions Incast Control TCP (ICTCP) and Data Center TCP (DCTCP). We also show that TCP PLATO can be implemented using commodity switches with Weighted Random Early Detection (WRED) function.

Original languageEnglish (US)
Article number140107
Pages (from-to)65-76
Number of pages12
JournalIEEE Journal on Selected Areas in Communications
Volume32
Issue number1
DOIs
StatePublished - Jan 2014

Fingerprint

Labeling
Switches
Packet loss
Labels
Simulators
Communication

Keywords

  • Data Center Networks
  • TCP Incast

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Computer Networks and Communications

Cite this

TCP PLATO : Packet labelling to alleviate time-out. / Shukla, Shikhar; Chan, Shingau; Tam, Adrian S W; Gupta, Abhishek; Xu, Yang; Chao, H. Jonathan.

In: IEEE Journal on Selected Areas in Communications, Vol. 32, No. 1, 140107, 01.2014, p. 65-76.

Research output: Contribution to journalArticle

Shukla, Shikhar ; Chan, Shingau ; Tam, Adrian S W ; Gupta, Abhishek ; Xu, Yang ; Chao, H. Jonathan. / TCP PLATO : Packet labelling to alleviate time-out. In: IEEE Journal on Selected Areas in Communications. 2014 ; Vol. 32, No. 1. pp. 65-76.
@article{86f9707497e1437bb774b75b3a73f3dd,
title = "TCP PLATO: Packet labelling to alleviate time-out",
abstract = "Many applications (e.g., cluster based storage and MapReduce) in modern data centers require a high fan-in, many-to-one type of data communication (known as TCP incast), which could cause severe incast congestion in switches and result in TCP goodput collapse, substantially degrading the application performance. The root cause of such a collapse is the long idle period of the Retransmission Timeout (RTO) that is triggered at one or more senders by packet losses in congested switches. In this paper we develop a packet labelling scheme PLATO, which improves the loss detection capabilities of NewReno using an innovative packet labelling system. Packets carrying this special label are preferentially enqueued, at the switch. This allows TCP to detect packet loss using three duplicate acknowledgements, instead of the time expensive RTO; thus avoiding the goodput collapse. PLATO makes minor modifications to NewReno and does not alter its congestion control mechanism. The implementation and simulations have been done in Network Simulator 3 (NS3). PLATO's performance is significantly better than NewReno as well as state-of-art incast solutions Incast Control TCP (ICTCP) and Data Center TCP (DCTCP). We also show that TCP PLATO can be implemented using commodity switches with Weighted Random Early Detection (WRED) function.",
keywords = "Data Center Networks, TCP Incast",
author = "Shikhar Shukla and Shingau Chan and Tam, {Adrian S W} and Abhishek Gupta and Yang Xu and Chao, {H. Jonathan}",
year = "2014",
month = "1",
doi = "10.1109/JSAC.2014.140107",
language = "English (US)",
volume = "32",
pages = "65--76",
journal = "IEEE Journal on Selected Areas in Communications",
issn = "0733-8716",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "1",

}

TY - JOUR

T1 - TCP PLATO

T2 - Packet labelling to alleviate time-out

AU - Shukla, Shikhar

AU - Chan, Shingau

AU - Tam, Adrian S W

AU - Gupta, Abhishek

AU - Xu, Yang

AU - Chao, H. Jonathan

PY - 2014/1

Y1 - 2014/1

N2 - Many applications (e.g., cluster based storage and MapReduce) in modern data centers require a high fan-in, many-to-one type of data communication (known as TCP incast), which could cause severe incast congestion in switches and result in TCP goodput collapse, substantially degrading the application performance. The root cause of such a collapse is the long idle period of the Retransmission Timeout (RTO) that is triggered at one or more senders by packet losses in congested switches. In this paper we develop a packet labelling scheme PLATO, which improves the loss detection capabilities of NewReno using an innovative packet labelling system. Packets carrying this special label are preferentially enqueued, at the switch. This allows TCP to detect packet loss using three duplicate acknowledgements, instead of the time expensive RTO; thus avoiding the goodput collapse. PLATO makes minor modifications to NewReno and does not alter its congestion control mechanism. The implementation and simulations have been done in Network Simulator 3 (NS3). PLATO's performance is significantly better than NewReno as well as state-of-art incast solutions Incast Control TCP (ICTCP) and Data Center TCP (DCTCP). We also show that TCP PLATO can be implemented using commodity switches with Weighted Random Early Detection (WRED) function.

AB - Many applications (e.g., cluster based storage and MapReduce) in modern data centers require a high fan-in, many-to-one type of data communication (known as TCP incast), which could cause severe incast congestion in switches and result in TCP goodput collapse, substantially degrading the application performance. The root cause of such a collapse is the long idle period of the Retransmission Timeout (RTO) that is triggered at one or more senders by packet losses in congested switches. In this paper we develop a packet labelling scheme PLATO, which improves the loss detection capabilities of NewReno using an innovative packet labelling system. Packets carrying this special label are preferentially enqueued, at the switch. This allows TCP to detect packet loss using three duplicate acknowledgements, instead of the time expensive RTO; thus avoiding the goodput collapse. PLATO makes minor modifications to NewReno and does not alter its congestion control mechanism. The implementation and simulations have been done in Network Simulator 3 (NS3). PLATO's performance is significantly better than NewReno as well as state-of-art incast solutions Incast Control TCP (ICTCP) and Data Center TCP (DCTCP). We also show that TCP PLATO can be implemented using commodity switches with Weighted Random Early Detection (WRED) function.

KW - Data Center Networks

KW - TCP Incast

UR - http://www.scopus.com/inward/record.url?scp=84891604726&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84891604726&partnerID=8YFLogxK

U2 - 10.1109/JSAC.2014.140107

DO - 10.1109/JSAC.2014.140107

M3 - Article

VL - 32

SP - 65

EP - 76

JO - IEEE Journal on Selected Areas in Communications

JF - IEEE Journal on Selected Areas in Communications

SN - 0733-8716

IS - 1

M1 - 140107

ER -