Understanding TCP incast throughput collapse in datacenter networks

Authors:
Yanpei Chen;Rean Griffith;Junda Liu;Randy H. Katz;Anthony D. Joseph
Affiliations:
RAD Lab EECS Dept UC Berkeley, Berkeley, CA, USA;RAD Lab EECS Dept UC Berkeley, Berkeley, CA, USA;RAD Lab EECS Dept UC Berkeley, Berkeley, CA, USA;RAD Lab EECS Dept UC Berkeley, Berkeley, CA, USA;Intel Labs Berkeley, Berkeley, CA, USA
Venue:
Proceedings of the 1st ACM workshop on Research on enterprise networking
Year:
2009

Citing 8
Cited 31

Congestion avoidance and control

SIGCOMM '88 Symposium proceedings on Communications architectures and protocols
Analysis of the increase and decrease algorithms for congestion avoidance in computer networks

Computer Networks and ISDN Systems
The Google file system

SOSP '03 Proceedings of the nineteenth ACM symposium on Operating systems principles
Design, deployment, and use of the DETER testbed

DETER Proceedings of the DETER Community Workshop on Cyber Security Experimentation and Test on DETER Community Workshop on Cyber Security Experimentation and Test 2007
MapReduce: simplified data processing on large clusters

Communications of the ACM - 50th anniversary issue: 1958 - 2008
Measurement and analysis of TCP throughput collapse in cluster-based storage systems

FAST'08 Proceedings of the 6th USENIX Conference on File and Storage Technologies
A scalable, commodity data center network architecture

Proceedings of the ACM SIGCOMM 2008 conference on Data communication
Safe and effective fine-grained TCP retransmissions for datacenter communication

Proceedings of the ACM SIGCOMM 2009 conference on Data communication

Safe and effective fine-grained TCP retransmissions for datacenter communication

Proceedings of the ACM SIGCOMM 2009 conference on Data communication
The nature of data center traffic: measurements & analysis

Proceedings of the 9th ACM SIGCOMM conference on Internet measurement conference
Data center TCP (DCTCP)

Proceedings of the ACM SIGCOMM 2010 conference
XCo: explicit coordination to prevent network fabric congestion in cloud computing cluster platforms

Proceedings of the 19th ACM International Symposium on High Performance Distributed Computing
Sources of instability in data center multicast

Proceedings of the 4th International Workshop on Large Scale Distributed Systems and Middleware
ICTCP: Incast Congestion Control for TCP in data center networks

Proceedings of the 6th International COnference
Efficiently measuring bandwidth at all time scales

Proceedings of the 8th USENIX conference on Networked systems design and implementation
Better never than late: meeting deadlines in datacenter networks

Proceedings of the ACM SIGCOMM 2011 conference
Managing data transfers in computer clusters with orchestra

Proceedings of the ACM SIGCOMM 2011 conference
Cross-layer flow and congestion control for datacenter networks

Proceedings of the 3rd Workshop on Data Center - Converged and Virtual Ethernet Switching
Purlieus: locality-aware resource allocation for MapReduce in a cloud

Proceedings of 2011 International Conference for High Performance Computing, Networking, Storage and Analysis
Practical TDMA for datacenter ethernet

Proceedings of the 7th ACM european conference on Computer Systems
Cutting MapReduce cost with spot market

HotCloud'11 Proceedings of the 3rd USENIX conference on Hot topics in cloud computing
OSA: an optical switching architecture for data center networks with unprecedented flexibility

NSDI'12 Proceedings of the 9th USENIX conference on Networked Systems Design and Implementation
The TCP outcast problem: exposing unfairness in data center networks

NSDI'12 Proceedings of the 9th USENIX conference on Networked Systems Design and Implementation
Preventing TCP incast throughput collapse at the initiation, continuation, and termination

Proceedings of the 2012 IEEE 20th International Workshop on Quality of Service
Explicit coordination to prevent congestion in data center networks

Cluster Computing
Deadline-aware datacenter tcp (D2TCP)

Proceedings of the ACM SIGCOMM 2012 conference on Applications, technologies, architectures, and protocols for computer communication
DeTail: reducing the flow completion time tail in datacenter networks

Proceedings of the ACM SIGCOMM 2012 conference on Applications, technologies, architectures, and protocols for computer communication
A switch-based approach to throughput collapse and starvation in data centers

Computer Networks: The International Journal of Computer and Telecommunications Networking
Deadline-aware datacenter tcp (D2TCP)

ACM SIGCOMM Computer Communication Review - Special october issue SIGCOMM '12
DeTail: reducing the flow completion time tail in datacenter networks

ACM SIGCOMM Computer Communication Review - Special october issue SIGCOMM '12
Rethinking end-to-end congestion control in software-defined networks

Proceedings of the 11th ACM Workshop on Hot Topics in Networks
Distributed adaptive routing for big-data applications running on data center networks

Proceedings of the eighth ACM/IEEE symposium on Architectures for networking and communications systems
Ensuring connectivity via data plane mechanisms

nsdi'13 Proceedings of the 10th USENIX conference on Networked Systems Design and Implementation
Got loss? Get zOVN!

Proceedings of the ACM SIGCOMM 2013 conference on SIGCOMM
R2D2: bufferless, switchless data center networks using commodity ethernet hardware

Proceedings of the ACM SIGCOMM 2013 conference on SIGCOMM
ICTCP: incast congestion control for TCP in data-center networks

IEEE/ACM Transactions on Networking (TON)
Efficient new delayed ACK for TCP: old problem, new insight

Proceedings of the 16th ACM international conference on Modeling, analysis & simulation of wireless and mobile systems
Review: A survey on architectures and energy efficiency in Data Center Networks

Computer Communications
IDTCP: An effective approach to mitigating the TCP incast problem in data center networks

Information Systems Frontiers

Quantified Score

Hi-index	0.00

Visualization

Abstract

TCP Throughput Collapse, also known as Incast, is a pathological behavior of TCP that results in gross under-utilization of link capacity in certain many-to-one communication patterns. This phenomenon has been observed by others in distributed storage, MapReduce and web-search workloads. In this paper we focus on understanding the dynamics of Incast. We use empirical data to reason about the dynamic system of simultaneously communicating TCP entities. We propose an analytical model to account for the observed Incast symptoms, identify contributory factors, and explore the efficacy of solutions proposed by us and by others.