Please use this identifier to cite or link to this item: http://arks.princeton.edu/ark:/88435/dsp015d86p039j
 Title: Copacetic: An Optimization Platform for Software Defined Networks Authors: Choi, Andrew Advisors: Rexford, Jennifer Contributors: Vanderbei, Robert Department: Operations Research and Financial Engineering Class Year: 2014 Abstract: Network management is a key part of any modern company's IT department. Software defined networking allows for better and more sophisticated control of the many functions that computer networks perform. In addition, these policies can be implemented on commodity hardware. Traditionally, specialized appliances were required to implement some critical network management roles, such as load balancing. The adoption of SDN techniques and deployments into the datacenter has allowed operators who would have otherwise been priced out of the market to adopt some of these functions. However, deploying a SDN-based solution is just one element of the ecosystem. It is easy to see that data about the underlying network can be used to improve upon the policies that are being used to manage the network, and that a framework for taking measurement data and incorporating it into the formulation of better policies would be useful to many network operators. Given this, we propose a design for a generalizable measure-optimize-implement cycle that will improve network management. Then, we introduce an implementation of the platform built on the Pyretic network controller, and examine the result from both a theoretical standpoint and a practical one. Finally, we suggest where the intersection of SDN and optimization research may continue. Extent: 67 URI: http://arks.princeton.edu/ark:/88435/dsp015d86p039j Type of Material: Princeton University Senior Theses Language: en_US Appears in Collections: Operations Research and Financial Engineering, 2000-2016

Files in This Item:
File SizeFormat