Let me rephrase that: Transparent optimization in SDNs

Santhosh Prabhu, Mo Dong, Tong Meng, P. Brighten Godfrey, Matthew Caesar

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Enterprise networks today have highly diverse correctness requirements and relatively common performance objectives. As a result, preferred abstractions for enterprise networks are those which allow matching correctness specification, while transparently managing performance. Existing SDN network management architectures, however, bundle correctness and performance as a single abstraction. We argue that this creates an SDN ecosystem that is unnecessarily hard to build, maintain and evolve. We advocate a separation of the diverse correctness abstractions from generic performance optimization, to enable easier evolution of SDN controllers and platforms. We propose Oreo, a first step towards a common and relatively transparent performance optimization layer for SDN. Oreo performs the optimization by first building a model that describes every flow in the network, and then performing network-wide, multi-objective optimization based on this model without disrupting higher level correctness.

Original languageEnglish (US)
Title of host publicationSOSR 2017 - Proceedings of the 2017 Symposium on SDN Research
PublisherAssociation for Computing Machinery, Inc
Pages41-47
Number of pages7
ISBN (Electronic)9781450349475
DOIs
StatePublished - Apr 3 2017
Event2017 Symposium on SDN Research, SOSR 2017 - Santa Clara, United States
Duration: Apr 3 2017Apr 4 2017

Publication series

NameSOSR 2017 - Proceedings of the 2017 Symposium on SDN Research

Other

Other2017 Symposium on SDN Research, SOSR 2017
CountryUnited States
CitySanta Clara
Period4/3/174/4/17

Keywords

  • Optimization
  • Software-defined networking

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Software

Fingerprint Dive into the research topics of 'Let me rephrase that: Transparent optimization in SDNs'. Together they form a unique fingerprint.

Cite this