doi:10.1016/j.peva.2007.06.013
Copyright © 2007 Elsevier Ltd All rights reserved.
Application-layer multipath data transfer via TCP: Schemes and performance tradeoffs
References and further reading may be available for this article. To view references and further reading you must
purchase this article.
Bing Wanga, 
, 
, Wei Weib, , Jim Kurosec, , Don Towsleyc, , Krishna R. Pattipatid, , Zheng Guoa, and Zheng Penga,
aComputer Science and Engineering Department, University of Connecticut, Storrs, CT 06269, United States
bUnited Technologies Research Center, East Hartford, CT 06108, United States
cComputer Science Department, University of Massachusetts, Amherst, MA 01003, United States
dElectrical and Computer Engineering Department, University of Connecticut, Storrs, CT 06269, United States
Available online 26 June 2007.
Abstract
For applications involving data transmission from multiple sources, an important problem is: when sources are allowed to use multiple paths, how does one select paths and control the sending rates on the paths to maximize the aggregate sending rate of the sources? We consider this problem in the context of an overlay network by allowing a source to send data over 
overlay paths to its destination. This problem is NP-hard, and we develop an iterative distributed heuristic to solve it. In each iteration, we first select paths and then control the sending rates on the multiple paths to maximize the aggregate sending rate of the sources. For rate control, we develop an application-level multipath rate controller via TCP. This controller is easy to deploy and maximizes the aggregate sending rate of the sources in certain settings. To the best of our knowledge, this is the first distributed application-level controller with such an optimality property. For path selection, we prove that the problem of optimal overlay path selection is NP-hard and propose randomized path-selection algorithms. Our performance evaluation demonstrates that our iterative heuristic performs very well in a wide range of settings. Furthermore, a small number of paths, 2–4, and a small amount of extra bandwidth in the network are sufficient to realize most of the performance gains.
Keywords: Multipath data transfer; Application-level rate control; Path selection
Fig. 1. Problem setting, k=2 in this example.
Fig. 2. Application-level multipath rate control: source s performs parameter-adjustment and rate-increment steps at the beginning of the nth control interval, n≥1, α and γ are constants, α>1,γ>1, 
s is a small positive constant.
Fig. 3. An example illustrating the rate adjustment by our rate controller. In the rate adjustment sequence, s=1 Mbps, βsj(0)=0.
Fig. 4. Evaluation of our rate controller for homogeneous source demands and k=2: (a) Results in a single run, using AIAD, b=0.5, f=1; (b) CDF of the relative difference of results from our controller and from CPLEX; (c) CCDF of the convergence time, =0.01.
(a) b=0
(b) b=0.5
(c) b=1.0
Fig. 7. Number of iterations required to find an optimal path selection, homogeneous source demands, k=2.
Corresponding author.
Abstract
Purchase PDF (1256 K)
E-mail Article
Add to my Quick Links
Cited By in Scopus (2)
Purchase PDF (406 K)
