Dissertation Defense Announcement
Dong Wei
Time: 10:00 AM, Thursday, April 15, 2004.
Place: Room 202, ECE Center, New Jersey Institute of Technology (NJIT), Newark, NJ.

"Virtual Private Network" (VPN) facilitates the communication among a set of sites and provides customers with predictable and secure network connections over a shared network infrastructure. Multiple sites of a private network may therefore communicate via the public infrastructure, mimicking the operation of the private network. The "classical" VPN service is realized on layer 2, by either Frame Relay (FR) or Asynchronous Transfer Mode (ATM). With FR or ATM, virtual circuits are created before traffic delivery. Since the bandwidth and buffers are reserved, the QoS requirements can be naturally guaranteed. In the past few years, layer 3 VPN technologies are widely deployed due to the desirable performance in terms of flexibility, scalability and simplicity. Layer 3 VPNs are built upon IP tunnels. Since IP is " best effort" in nature, the QoS requirement cannot be guaranteed by layer 3 VPNs. Actually, layer 3 VPN service can only provide secure connectivity, i.e., isolating each customer’s traffic from the others. With more applications on voice, audio and video being used in the Internet, the provision of QoS is no doubt one of the most critical considerations of the emerging services provided by ISPs. An intriguing question is: “Can we obtain the best of both layers 2 and 3 VPN? Can we provide guaranteed or predictable QoS, as in layer 2 VPNs, while maintaining the flexibility and simplicity in layer 3 VPN?" This question is the starting point of this study.

The recently proposed hose model for VPN possesses desirable properties in terms of flexibility, scalability and multiplexing gain. However, the "classic" fair bandwidth allocation schemes and weighted fair queuing schemes raise the issue of low overall utilization in this model. We propose a new fluid model for provider-provisioned VPN (PPVPN). Based on the proposed model, we develop an idealized fluid bandwidth allocation scheme, which is proven, analytically, to have the following properties: 1) maximizes the overall throughput of the VPN without compromising fairness; 2) provides a mechanism that enables the VPN customers to allocate the bandwidth according to their requirements by assigning different weights to different hose flows, and thus achieves the predictable QoS performance; and 3) improves the overall throughput of the ISPs' network.

To approximate the idealized fluid scheme, we develop a 2-dimensional deficit round robin (2-D DRR and 2-D DRR+) schemes, and a more scalable non-per-flow-based scheme for output queued switches. In order to approximate H-GPS, 2-D DRR and 2-D DRR+ schemes can be extended to multiple dimensions and deployed in a “tiered” scheduling architecture to meet the requirement for flexible and accurate bandwidth allocation. The performances of these schemes are analyzed. We also discuss how to employ our proposed schemes in the current layer 3 VPN technologies.

Committee Members:

Professor Nirwan Ansari, the ECE department of NJIT (Advisor)

Professor Symeon Papavassiliou, the ECE department of NJIT

Professor Lev Zakrevski, the ECE department of NJIT

Professor Teunis Ott, the CIS department of NJIT

Dr. Jianguo Chen, Agere Systems