Ph.D. Thesis Proposal by Hyojoon Kim

Event Details
  • Date/Time:
    • Wednesday September 3, 2014 - Thursday September 4, 2014
      1:00 pm - 2:59 pm
  • Location: KACB 2100
  • Phone:
  • URL:
  • Email:
  • Fee(s):
  • Extras:
No contact information submitted.

Summary Sentence: Facilitating Network Management with Software Defined Networking

Full Summary: No summary paragraph submitted.

Ph.D. Thesis Proposal Announcement

Title:  Facilitating Network Management with Software Defined Networking

 Hyojoon Kim
School of Computer ScienceCollege of Computing
Georgia Institute of Technology

Date: September 3rd, 2014 (Wednesday)
Time: 1pm--3pm ET
Location: KACB 2100

Committee:Dr. Nick Feamster (Advisor), School of Computer Science, Georgia TechDr. Ellen Zegura, School of Computer Science, Georgia TechDr. Mostafa Ammar, School of Computer Science, Georgia TechDr. Nate Foster, Department of Computer Science, Cornell UniversityDr. Yoshio Turner, Networking and Mobility Lab, HP Labs-Palo Alto Abstract:

Network management refers to activities, procedures, and methods for operating, maintaining, and provisioning networked systems. Network operators are responsible of providing reliable and secure communication between end hosts as well as to and from the Internet. However, managing and configuring a network is hard. Complexity of the task causes significant amount of misconfiguration that disrupt day-to-day network operation. Yet, we understand very little about the nature of network configuration while such knowledge is essential for building a better solution.

This thesis proposal focuses on two closely related fields: (1) analysis of conventional network management methods to better understand the problem, and (2) implementation of better network management systems with Software Defined Networking (SDN). First, I present analytical studies that help understand how networks are configured and managed. I analyze over five years of historical network configuration files from two big campus networks. Our study shows that a network can experience a lot of changes; number of lines that change in all configuration files ranges from 200,000 to 800,000 lines per year. Second, based on the findings from the analysis study, I present two distinct solutions that are both based on SDN. Kinetic is a SDN-based controller platform that helps simplify network management. Kinetic supports expressing event-driven, dynamic network policies as finite state machines, and can automatically reconfigure the network by reacting to various types of events. CoroNet is a SDN-based service that provides automated recovery in face of a special kind of (but common) network event: data-plane failure such as switch and link failures. 

Additional Information

In Campus Calendar

Graduate Studies

Invited Audience
cs, Graduate Student, thesis
  • Created By: Danielle Ramirez
  • Workflow Status: Published
  • Created On: Aug 28, 2014 - 7:42am
  • Last Updated: Oct 7, 2016 - 10:08pm