Abstract:
network buffering, cross traffic, and socket buffer auto-sizing Abstract — TCP is often blamed that it cannot use efficiently network paths with high Bandwidth-Delay Product (BDP). The BDP is of fundamental importance because it determines the required socket buffer size for maximum throughput. In this paper, we re-examine the BDP concept, considering the effects of network buffering and cross traffic on the ‘bandwidth ’ and ‘delay ’ characteristics of a path. We show that, with careful socket buffer sizing, a bulk TCP transfer can saturate a network path independent of the BDP or the available network buffers. In a non-congested path, there is a certain socket buffer size (which depends on the cross traffic type) that maximizes the throughput of a bulk TCP transfer. In a congested path, the TCP throughput is maximized when the connection is limited by the congestion window, rather than by the socket buffers. Finally, we present an application-layer mechanism (SOBAS) that automatically adjusts the socket buffer size close to its optimal value, based on direct measurements of the maximum received throughput and of the round-trip time, without requiring prior knowledge of the path characteristics.
Citations
|
827
|
Modeling TCP Throughput: a Simple Model and its Empirical Validation
– Padhye, Firoiu, et al.
- 1998
|
|
607
|
End-to-end Internet Packet Dynamics
– Paxson
- 1999
|
|
273
|
Requirements for Internet Hosts – Communication Layers
– Braden
- 1989
|
|
221
|
Computer Networks: A Systems Approach
– Peterson, Davie
- 1996
|
|
184
|
End-to-end available bandwidth: Measurement methodolody, dynamics, and relation with tcp throughput
– JAIN, DOVROLIS
- 2002
|
|
179
|
Congestion Control for High Bandwidth-Delay Product Networks
– Katabi, Handley, et al.
- 2002
|
|
160
|
On Estimating End-to-End Network Path Properties
– Allman
- 1999
|
|
110
|
HighSpeed TCP for Large Congestion Windows
– Floyd
- 2003
|
|
99
|
Automatic TCP Buffer Tuning
– Semke, Mahdavi, et al.
- 1998
|
|
77
|
A New End-to-End Probing and Analysis Method for Estimating Bandwidth Bottlenecks
– Melander, Bjorkman, et al.
- 2000
|
|
76
|
PSockets: The Case for Application-level Network Striping for Data Intensive Applications using High Speed Wide Area Networks
– Sivakumar, Mazzucco, et al.
|
|
48
|
The end-to-end performance effects of parallel TCP sockets on a lossy wide-area network
– Hacker, Athey, et al.
- 2001
|
|
42
|
TCP Extensions for High Performance
– Jacobson, Borman
- 1992
|
|
27
|
Limited Slow-Start for TCP with Large Congestion Windows
– Floyd
- 2004
|
|
27
|
A framework for defining empirical bulk transfer capacity metrics, rfc3148
– MATHIS, ALLMAN
- 2001
|
|
25
|
A TCP Tuning Daemon
– Dunigan, Mathis, et al.
- 2002
|
|
25
|
Measuring Bottleneck Link Speed
– Carter, Crovella
- 1996
|
|
24
|
Measuring End-to-End Bulk Transfer Capacity
– Allman
- 2001
|
|
24
|
A Poisson Limit for Buffer Overflow Probabilities
– Ramanan, Cao
- 2002
|
|
14
|
A Swifter Start for TCP
– Partridge, Rockwell, et al.
- 2002
|
|
12
|
Stochastic models for throughput analysis of randomly arriving elastic flows
– Kherani, Kumar
|
|
11
|
Automatic TCP socket buffer tuning
– Lui, Ferguson
|
|
9
|
A Comparison of TCP Automatic Tuning Techniques for Distributed Computing
– Weigle, Feng
- 2002
|
|
8
|
Internet End-to-End Performance Monitoring for the High-Energy Nuclear and Particle Physics Community
– Matthews, Cottrell
- 2000
|
|
5
|
Dynamic Right-Sizing in FTP (drsFTP): Enhancing Grid Performance in User-Space
– Gardner, Feng, et al.
- 2002
|
|
4
|
Enabling High Performance Data Transfers
– Mathis, Reddy
- 2003
|
|
4
|
Network Characterization Service.” http://wwwdidc.lbl.gov/pipechar
– Guojun
- 2001
|
|
3
|
TCP Tuning Guide for
– Tierney
- 2001
|
|
2
|
Is TCP an Adequate Protocol for High-Performance Computing Needs
– Feng
- 2000
|
|
1
|
Kuzmanovic and E.W.Knightly, “TCP-LP: A Distributed Algorithm for Low Priority Data Transfer
– unknown authors
- 2003
|
