Research
My general research interests include finding practical and efficient solutions for
routing, transport, and security in wireless MANETs and DTNs.
BBN Technologies
Coming soon...
WINLAB, Rutgers University - MobilityFirst
While at WINLAB, Rutgers University, I helped lead the local routing effort for the MobilityFirst FIA project.
This three-year NSF project is aimed at a clean-slate
design and validation of a mobility-centric Internet architecture.
MobilityFirst is founded on the premise that mobile, wireless communication
is replacing the historic fixed-host/server model, and hence the future Internet
architecture must provide inherent support for these mobile devices. In
particular, we developed global and local routing protocols
that met the many new challenges of a mobility-centric Internet, such as link
quality fluctuation, a large variance in connectivity levels, and
security/privacy issues.
University of Illinois - Flexible and Robust DTN Routing
Many DTNs exhibit human-centric behavior, in that mobility and communication
patterns tend to following human-based social interactions. Therefore,
group-based communication is a very natural and useful form of communication in
DTNs. To this end, I have developed a series of routing techniques and
protocols to enable one-to-many forms of communication, such as anycast,
manycast, and multicast. Furthermore, due to the inherently untrustworthy
environment found in DTNs along with the high probability of disconnection from
a centralized trust authority, I have developed a robust group management
protocol that is accurate even in the face of multiple attacks and runs without
the need for cryptography. Two interesting works from this series can be found
in the ACM CHANTS 2010 proceedings. The most recent work from this project,
exploring the space between anycast and multicast, can be found as a UIUC
technical report here.
University of Illinois - Phoenix Project / Efficient DTN Routing
Disruption-tolerant networking allows communication between mobile ad-hoc
devices without reliance on static infrastructure, such as cellular towers and
access points. By utilizing store-carry-and-forward techniques, along with smart
replication, DTNs are able to route data through highly varying levels of
connectivity and partitioning. As part of the Phoenix Project, I
developed a highly efficient DTN unicast routing protocol, called
Encounter-Based Routing (EBR). EBR, found in the IEEE INFOCOM 2009
proceedings, achieves two objectives that are
seemingly at odds with one another: (1) resource-friendliness and (2) high
message delivery. By learning and utilizing mobility structure, EBR is able to
heavily limit replication while at the same time getting replicas to high-value
nodes.
|
