• GMPLS protocol development: Specifically, we are developing new ideas that can improve the performance of fault management techniques in GMPLS. We are currently discussing with our collaborations at Alcatel for possible extension of the research, and, hopefully, turn the research results into industry standards.
• Quality of service based techniques in traffic provisioning and restoration: Most existing protection and restoration schemes assume only one class of service. Customers' traffic in a WDM network is either protected or not. However, when different preferences can be given to different connections, new network design models are needed. With our simulation tool, it is possible to start from understanding the effects of different QoS policies on network performances. Then, from that on, design appropriate QoS policies for the next generation WDM networks.
• With the rapid deployment of optical switching technology in layer 2, service providers are given new potentials to develop new services. For instance, a provider may want to provide "very short duration" connections to some special costumers (say a financial institute who wants to backup its data at midnight everyday). How much service disruption will these services introduce to normal network operation, and to what extend will these services destabilize the routing protocols. These are the problems that are impossible to study without a detailed network simulation tool. We have the advantage to be a lead in this area.
• IP-backbone networks are intrinsically different from telephone networks. However, network researchers only have limited vision on what these differences are and how the next generation network architecture should adopt. This is understandable because of the short history of large-scale packet-switched network. With our extensive knowledge on both IP and optical networking, we have the potential to help bring some highly impact results to light.

For our publications, feel free to visit our Publications page.