Systems and Networking. My research has involved the design and analysis of protocols and architecture in the following areas:

• Mobile and Ubiquitous Computing: Ad Hoc and Sensor Networks, Disruption Tolerant Networks

Disruption Tolerant Networks (DTNs) are wireless mobile networks in which nodes frequently lose their connections to others, yet have routing paths over time. These networks occur in many important applications such as in sparsely deployed sensor networks to track animals, vehicular networks, temporary networks in disastrous sites, and military situations. Compared to a traditional mobile ad hoc network (MANET) in which nodes form a connected network, DTNs raise many new challenges due to the frequent disruptions. For example, in both network paradigms, many nodes depend on untethered devices with limited energy supplies. Thus, careful power management is necessary to keep the networks functional for a long time. However, power management mechanisms designed for MANETs are not appropriate for use in DTNs because they were designed under assumption that nodes always have other nodes in the neighborhood. In DTNs, nodes can save significant additional energy without degrading network performance by putting their wireless interfaces to sleep when they are isolated. The major challenge is balancing sleeping periods with wake-up periods of wireless interfaces, so that valuable and infrequent communication opportunities between nodes are well utilized. In my doctoral research, I have focused on the power management issues in DTNs.

• Hyewon Jun, Wenrui Zhao, Mostafa Ammar, Ellen Zegura, and Chungki Lee, 

Trading Latency for Energy in Densely Deployed Wireless Ad Hoc Networks using Message Ferrying, 

In Elsevier Journal of Ad Hoc Networks (to appear), 2006. 

A preliminary version appeared in IEEE PerCom International Workshop on Pervasive Wireless Networking (PWN), 2005 (pdf). 

• Hyewon Jun, Mostafa Ammar, Mark Corner, and Ellen Zegura, 

Hierarchical Power Management in Disruption Tolerant Networks with Traffic-Aware Optimization, 

In ACM SIGCOMM Workshop on Challenged Networks (CHANTS), September, 2006 

• Hyewon Jun, Mostafa H. Ammar, and Ellen W. Zegura, 

Power Management in Delay Tolerant Networks: A Framework and Knowledge-Based Mechanisms, 

In IEEE Communications Society Conference on Sensor and Ad Hoc Communications and Networks (SECON), September, 2005

The IP Multimedia Subsystem (IMS) is the next-generation standard for telecommunication networks that remodels cellular network systems onto the IP network platform. The aim of IMS is to provide all multimedia services such as Voice-over-IP in such a way that network service providers can control and charge for each service while taking advantage of successful IP technologies. To provide flexible session control, IMS uses the Session Initiation Protocol (SIP) to set up and tear down multimedia sessions through a potentially large chain of proxies. These SIP proxies may perform many computation intensive jobs such as keeping track of transaction among proxies and filtering/matching application options. While these jobs provide advantages, they may limit the capacity of a system. To evaluate their impact on the whole system, a tool is necessary. However, given the system was under development, tools to test various architectural options were unavailable. Thus, we simplified the system model in the system perspective and developed a simulation tool to evaluate architectural choices and algorithms to process calls. We also proposed and evaluated a stateless core approach in which proxies in the edge of systems keep the transaction states, while the proxies in the center waive the jobs depending on the load in the systems.

• Mauricio Cortes, Jairo Esteban, and Hyewon Jun, 

Diabelli: an IMS Simulation Tool, 

In Bell Labs Technical Journal, Vol. 10, No. 4, 2006 

• Mauricio Cortes, Jairo Esteban, and Hyewon Jun, 

Towards Stateless Core: Improving SIP Proxy Scalability,

In IEEE Globecom Conference, November, 2006

Much effort has been dedicated to the development of application-layer and network-layer mechanisms for selecting the "best" server among a set of replicated servers. However, when to invoke and use a server selection mechanism lies on clients' choice. We use the term "binding" to refer to the function that queries a server selection mechanism and makes decisions about how to use the results. We consider two extreme cases of binding frequency, Initial Binding at the start of a connection, and Continuous Binding for each packet. We designed a seamless connection migration scheme to use a Continuous Binding scheme, and then compare the two extremes to one another.

• Hyewon Jun, Matt Sanders, Mostafa Ammar, and Ellen Zegura, 

Binding Clients to Replicated Servers: Initial and Continuous Binding, 

In IEEE International Workshop on Future Trends of Distributed Computing Systems (FTDCS), 2003. 

A longer version of the paper appeared in Technical Report GIT-CC-03-09, Georgia Institute of Technology, 2003 (pdf).