Expression and Composition of Optimization-Based Applications for Software-Defined Networking Public Deposited

Downloadable Content

Download PDF
Last Modified
  • March 19, 2019
  • Heorhiadi, Victor
    • Affiliation: College of Arts and Sciences, Department of Computer Science
  • Motivated by the adoption of the Software Defined Networking and its increasing focus on applications for resource management, we propose a novel framework for expressing network optimization applications. Named the SDN Optimization Layer (SOL), the framework and its extensions alleviate the burden of constructing optimization applications by abstracting the low-level details of mathematical optimization techniques such as linear programming. SOL utilizes the path abstraction to express a wide variety of network constraints and resource-management logic. We show that the framework is general and efficient enough to support various classes of applications. We extend SOL to support composition of multiple applications in a fair and resource-efficient way. We demonstrate that SOL’s composition produces better resource efficiency than previously available composition approaches and is tolerant to network variations. Finally, as a case study, we develop a new application for load balancing network intrusion prevention systems, called SNIPS. We highlight the challenges in developing the SNIPS optimization from the ground up, show SOL’s (conceptually) simplified version, and verify that both produce nearly identical solutions.
Date of publication
Resource type
Rights statement
  • In Copyright
  • Benson, Theophilus
  • Sekar, Vyas
  • Reiter, Michael
  • Aikat, Jay
  • Gupta, Anupam
  • Doctor of Philosophy
Degree granting institution
  • University of North Carolina at Chapel Hill Graduate School
Graduation year
  • 2017

This work has no parents.