Revisiting Flow-Based Load Balancing: Stateless Path Selection in Data Center Networks

Mon, 12/24/2012 - 11:50 by Gregory Detal


Hash-based load-balancing techniques are widely used to distribute the load over multiple forwarding paths and preserve the packet sequence of transport-level flows. Forcing a long-lived, i.e., elephant, flow to follow a specific path in the network is a desired mechanism in data center networks to avoid crossing hot spots. This limits the formation of bottlenecks and so improves the network use. Unfortunately, current per-flow load-balancing methods do not allow sources to deterministically force a specific path for a flow.

In this paper, we propose a deterministic approach enabling end hosts to steer their flows over any desired load-balanced path without relying on any packet header extension. By using an invertible mechanism instead of solely relying on a hash function in routers, our method allows to easily select the packet’s header field values in order to force the selection of a given load-balanced path without storing any state in routers.

We perform various simulations and experiments to evaluate the performance and prove the feasibility of our method using a Linux kernel implementation. Furthermore, we demonstrate with simulations and lab experiments how MultiPath TCP can benefit from the combination of our solution with a flow scheduling system that efficiently distributes elephant flows in large data center networks.

Gregory Detal, Christoph Paasch, Simon van der Linden, Pascal Mérindol, Gildas Avoine and Olivier Bonaventure
Computer Networks, 57(5):1204–1216, April 2013.
Full text
pdf    (750.26 KB)
Cite it
See here

IEEE Copyright Notice: This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. In most cases, these works may not be reposted without the explicit permission of the copyright holder.

ACM Copyright Notice: Copyright 1999 by the Association for Computing Machinery, Inc. Permission to make digital or hard copies of part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page or intial screen of the document. Copyrights for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, to republish, to post on servers, or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from Publications Dept., ACM Inc., fax +1 (212) 869-0481, or

Springer-Verlag LNCS Copyright Notice: The copyright of these contributions has been transferred to Springer-Verlag Berlin Heidelberg New York. The copyright transfer covers the exclusive right to reproduce and distribute the contribution, including reprints, translations, photographic reproductions, microform, electronic form (offline, online), or any other reproductions of similar nature. Online available from Springer-Verlag LNCS series.