Research Profile
Dr. Deng Pan’s research interests are generally in high speed computer networks, and his recent research focuses on resource and energy efficiency in data center networks.
The ongoing projects are as follows.
- Management of multiple deterministic and stochastic resources in data centers: A data center manages many different types of resources. Demands for some resources are deterministic, such as memory and storage, but demands for other resources are time varying and demonstrate a stochastic nature. We propose novel virtual machine placement algorithms to efficiently allocate the multiple resources and minimize the required number of servers. In particular, we use the probabilistic service availability defined in the service level agreement to provide service guarantees.
- Load-balanced multipath routing for data center networks: Data center contains hundreds of thousands of servers. In order to interconnect such a huge number of servers, modern data center networks adopt novel multi-rooted topologies for large bisection bandwidth. Such topologies offer multipath capability, i.e. multiple paths existing between the same pair of hosts. However, exiting routing protocols for traditional networks consider only a single path. We propose load-balanced multipath routing algorithms to fully explore the available bandwidth provided by the multi-rooted topologies, and maximize the admissible amount of traffic without generating hot spots.
- Joint host-network energy optimization for data centers: Data centers consume significant amounts of energy. As severs become more energy efficient with various energy saving techniques, the data center network has been accounting for 20% or more of the energy consumed by the entire data center. Although there exist a number of energy optimization solutions for data centers, they consider only either the hosts or network, but not both. We propose joint optimization schemes that simultaneously optimize virtual machine placement and network flow routing to maximize energy savings, and we also build an OpenFlow based prototype to experimentally demonstrate the effectiveness of our design.
Previous Accomplishments
- Collaboration
- Center of Excellence in Wireless and Information Technology, Stony Brook University
- Florida State University
- Tsinghua University, China
Relevant Publications
- Y. Li and Deng Pan, “OpenFlow based load balancing for fat-tree networks with multipath support,” IEEE International Conference on Communications (ICC), Budapest, Hungary, Jun. 2013.
- H. Jin, T. Cheocherngngarn, D. Levy, A. Smith, D. Pan, and Niki Pissinou, “Joint host-network optimization for energy-efficient data center networking,” IEEE International Parallel and Distributed Processing Symposium (IPDPS), Boston, MA, May 2013.
- H. Jin, D. Pan, J. Xu, and N. Pissinou, “Efficient VM placement with multiple deterministic and stochastic resources in data centers,” IEEE Global Communications Conference (GLOBECOM), Anaheim, CA, Dec. 2012.
- H. Jin, D. Pan, J. Liu, and N. Pissinou, “OpenFlow based flow level bandwidth provisioning for CICQ switches,” IEEE Transactions on Computers, accepted for publication.
- D. Pan and Y. Yang, “Flow based performance guarantee scheduling in buffered crossbar switches,” IEEE Transactions on Communications, accepted for publication.