Deng Pan

Associate Professor | FIU SCIS

Lecture Information:
  • September 22, 2017
  • 2:00 PM
  • ECS 241

Speaker Bio

(The talk was previous scheduled on September 8 but got postponed due to Hurricane Irma.)

Deng Pan received the BS and MS degrees in Computer Science from Xi’an Jiaotong University, China, in 1999 and 2002, respectively, and the PhD degree in Computer Science from State University of New York at Stony Brook in 2007. He is currently an Associate Professor in the School of Computing and Information Sciences, Florida International University. His recent research interests focus on network function virtualization, software-defined networking, and data center networking. He has published over fifty peer-reviewed papers in top journals and conferences, including the IEEE Transactions on Communications, IEEE Transactions on Computers, IEEE Transactions Network and Service Management, INFOCOM, and IPDPS. He has served in the Technical Program Committee and Local Arrangement Committee of a number of international conferences, including INFOCOM, GLOBECOM, and ICPP.

Description

Network Function Virtualization (NFV) revolutionizes the implementation of network functions, or middleboxes, as virtual machines that can run any on standard server. However, compared with traditional hardware-based network appliances, provisioning of network functions by NFV faces several challenges, including availability of multiple hosting servers, traffic changing effects of, and dependency between middleboxes. In this talk, we discuss the optimal placement of NFV middleboxes, and present solutions for middleboxes of different traffic changing effects and with different dependency relations. We start by showing the hardness of the studied problem, and then propose optimal algorithms for special cases and efficient heuristics for general scenarios. Further, we describe our implementation of the proposed solutions in a Software-Defined Networking (SDN) based prototype, and present extensive experimental and simulation results to demonstrate the effectiveness of our design.