August 31, 2018 –
Title: QoS in SDN-Based Large-Scale Networks
Ph.D. Candidate: Haitham Ghalwash
Major Advisor: Dr. Chun-Hsi Huang
Associate Advisors: Dr. Reda A. Ammar, Dr. Sanguthevar Rajasekaran
Day/Time: Friday, August 31, 2018 9:00am
Location: HBL Video Theater 2
Traditional networks often suffer from inefficient management and hard QoS guarantee. Software-Defined Networking (SDN) technology is the new promising approach providing abstraction and programmability of the network, which facilitates the efficient resolution of many security, performance, management and QoS issues in today’s networks. In this research, we leverage the capability of SDN to provide a level of QoS for network applications. We start by investigating the potential of deploying SDN in large-scale systems. We investigate the SDN-based network behavior and performance of two widely used topologies in large-scale data centers, namely Fat-Tree and BCube, using different network scales, traffic loads and traffic patterns. We then investigate the potential of supporting Big-Data applications using SDN. Preliminary experiments used a Hadoop multi-node cluster, installed in Docker containers, that is connected through a Fat-tree topology of OpenFlow switches. Packet forwarding is either by way of an SDN controller or the normal packet switching rules. We investigate the network performance in terms of throughput and execution time for the read/write and sorting operations. Finally, we propose a new approach to leverage SDN capability of forwarding and metric retrievals. The proposed Framework provides a guarantee QoS using either traffic classification or active/passive metrics-based route selection. The behavior and performance of different types of traffic, namely, UDP, TCP, VOIP, and traffic for a Big-data application are investigated. The monitoring module, built on top of the controller, provides port statistics, analyzes port utilization and probes the link delays. The route determination module employs a shortest path algorithm, with or without a QoS guarantee. Preliminary experimental results, on average, demonstrated a promising reduction rate in delay, jitter, packet loss, as well as port utilization.