By: Li Zhu
Advisor: Dr. Nirwan Ansari
Department of Electrical and Computer Engineering

Time: 1:00 PM, Thursday, April 28th, 2005
Place: Room 202, ECE Center, New Jersey Institute of Technology (NJIT), Newark NJ. Directions

Abstract

Differentiated Services (DiffServ) architecture is a promising means of providing QoS in Internet. In DiffServ networks, user flows are aggregated into a small number of service classes, or Per-hop Behaviors (PHBs), and DiffServ networks provide the same local per-node QoS to user traffics in the same PHB. Currently, three PHBs have been defined: Expedited Forwarding (EF), Assured Forwarding (AF), and Best Effort (BE).

Despite the existing extensive studies, few works have paid attention to the performance of DiffServ networks at the network level. Most of the available research works in the context of DiffServ networks focus on a very simple network configuration, i.e., a network with a single bottleneck link. However, a real network may have multiple bottleneck links and the research conclusions drawn from the single bottleneck link case may not be applicable to the much more generalized case. It is critical to understand the end-to-end QoS in DiffServ Networks both qualitatively and quantitatively, since end-to-end QoS is what users really concern. This dissertation aims to address these important issues.

EF PHB is supposed to provide low loss ratio and small delay services. It has been shown in literature that the end-to-end delay of the EF traffic can go to infinity even in an over-provisioned network. In this dissertation, a simple scalable aggregate scheduling scheme, called Youngest Serve First (YSF) algorithm is proposed. YSF is not only able to guarantee finite end-to-end delay, but also to keep a low scheduling complexity independent of the number of flows in the network.

With respect to the Best Effort service, first, Random Exponential Marking (REM), an existing AQM scheme, is studied under a new continuous time model, and its local stable condition is presented. Next, a novel virtual queue and rate based AQM scheme (VQR) is proposed, and its local stability is analyzed. Then, a new AQM framework, Edge-based AQM (EAQM) is proposed. EAQM is only deployed at the edge routers of the network and it achieves similar or better performance than traditional AQM schemes.

With respect to Assured Forwarding, a network-assist packet marking (NPM) scheme has been proposed to fairly allocate bandwidth among AF aggregates based on their Committed Information Rates (CIRs). By both theoretical analysis and experimental evaluation, it has been demonstrated that NPM can fairly distribute bandwidth among these aggregates in both single and multiple bottleneck link networks.

HighSpeed TCP (HSTCP) can provide reliable data transmission at very high speed over Internet. At the same time, Optical Burst Switching (OBS) becomes a promising technology to support future Internet backbone. A simple model for a single HSTCP connection over an OBS network is proposed. Based on this model, the impact of OBS on the throughput of a single HSTCP connection has been investigated.

Finally, current ongoing research works and future research directions are discussed.

Committee Members:

Dr. N. Ansari, Professor, ECE, NJIT, Dissertation Advisor
Dr. A. Kolarov, Technical leader, Cisco Systems, Inc., Committee Member
Dr. R. Rojas-Cessa, Assistant Professor, ECE, NJIT, Committee Member
Dr. S. Tekinay, Associate Professor, ECE, NJIT, Committee Member
Dr. M. Zhou, Professor, ECE, NJIT, Committee Member