Xpander: Towards optimal-performance datacenters

Asaf Valadarsky, Gal Shahaf, Michael Dinitz, Michael Schapira

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

70 Scopus citations

Abstract

Despite extensive efforts to meet ever-growing demands, today's datacenters often exhibit far-from-optimal performance in terms of network utilization, resiliency to failures, cost efficiency, incremental expandability, and more. Consequently, many novel architectures for high performance datacenters have been proposed. We show that the benefits of state-ofthe- art proposals are, in fact, derived from the fact that they are (implicitly) utilizing "expander graphs" (aka expanders) as their network topologies, thus unveiling a unifying theme of these proposals. We observe, however, that these proposals are not optimal with respect to performance, do not scale, or suffer from seemingly insurmountable deployment challenges. We leverage these insights to present Xpander, a novel datacenter architecture that achieves near-optimal performance and provides a tangible alternative to existing datacenter designs. Xpander's design turns ideas from the rich graph-theoretic literature on constructing optimal expanders into an operational reality. We evaluate Xpander via theoretical analyses, extensive simulations, experiments with a network emulator, and an implementation on an SDN-capable network testbed. Our results demonstrate that Xpander significantly outperforms both traditional and proposed datacenter designs. We discuss challenges to real-world deployment and explain how these can be resolved.

Original languageAmerican English
Title of host publicationCoNEXT 2016 - Proceedings of the 12th International Conference on Emerging Networking EXperiments and Technologies
PublisherAssociation for Computing Machinery, Inc
Pages205-219
Number of pages15
ISBN (Electronic)9781450342926
DOIs
StatePublished - 6 Dec 2016
Event12th ACM Conference on Emerging Networking Experiments and Technologies, ACM CoNEXT 2016 - Irvine, United States
Duration: 12 Dec 201615 Dec 2016

Publication series

NameCoNEXT 2016 - Proceedings of the 12th International Conference on Emerging Networking EXperiments and Technologies

Conference

Conference12th ACM Conference on Emerging Networking Experiments and Technologies, ACM CoNEXT 2016
Country/TerritoryUnited States
CityIrvine
Period12/12/1615/12/16

Bibliographical note

Funding Information:
We thank Nati Linial for many insightful conversations about expander constructions and Daniel Bienstock for his help in optimizing linear programs for throughput computation. We thank Brighten Godfrey and Ankit Singla for sharing the Jellyfish code with us and for helpful discussions. We thank Soudeh Ghorbani for helping with running experiments on the OCEAN platform. We thank Noga Alon, Alex Lubotzky, and Robert Krauthgamer for useful discussions about expanders. We also thank Jonathan Perry for suggesting Xpander's color coding, presented in Section 7. Finally, we thank our shepherd, Siddhartha Sen, and the anonymous CoNEXT reviewers, for their valuable feedback. The 1st author is supperted by a Microsoft Research Ph.D. Scholarship. The 2nd and 4th authors are supported by the Israeli Center for Research Excellence (I-CORE) in Algorithms. The 1st and 4th authors are supported by the PetaCloud industry-academia consortium. The third author is supported in part by NSF awards 1464239 and 1535887

Fingerprint

Dive into the research topics of 'Xpander: Towards optimal-performance datacenters'. Together they form a unique fingerprint.

Cite this