Research
Middleware for Service Oriented Computing in Ad Hoc Networks
The increasing ubiquity of wireless mobile computing platforms
has opened up the potential for unprecedented levels of
communication, coordination and collaboration among mobile
computing devices, most of which will occur in an ad hoc, on demand
manner. My research seeks to develop a middleware
supporting service provision in ad-hoc wireless networks. The
aim is to provide the software resources on mobile
devices that facilitate electronic collaboration. This is achieved
by applying the principles of service oriented computing (SOC),
an emerging paradigm that has seen success in wired settings.
The middleware is an adaptation and extension of the Jini model of SOC
to ad-hoc networks. The key contributions are (1)
a completely decentralized service advertisement and request
system that is geared towards handling the unpredictability and
dynamism of mobile ad-hoc networks, (2) an automated code
management system that can fetch, use and dispose of binaries
on an on-demand basis, (3) a mechanism supporting the logical
mobility of services, (4) an upgrade mechanism to extend the life
cycle of services, and (5) a lightweight security model that secures
all interactions, which is essential in an open environment.
Algorithms for Routing in Mobile Ad Hoc Networks
Much of the work on networking and communications is based on the
premise that components interact in one of two ways: either they are connected via
a stable wired or wireless network, or they make use of persistent storage repositories
accessible to the communicating parties. A new generation of networks raises serious
questions about the validity of these fundamental assumptions. In mobile ad hoc
wireless networks connections are transient and availability of persistent storage is rare.
Ad hoc routing algorithms focus on discovering routes between mobile hosts, given
that the configuration of the network changes continuously and the life time of
communication paths is extremely short. I am particularly concerned with achieving
communication among mobile devices that may
never find themselves in direct or indirect contact with each other at any point in time.
A unique aspect of my research is the idea of exploiting information associated
with the motion and availability profiles of the devices making up the ad hoc network.
The core idea is that two hosts could exchange messages if a third host (carrier)
moves from the sender to the receiver and can carry the message without establishing a
fully connected path from source to destination at any single moment.
This is the starting point for an investigation into a range of possible solutions whose
essential features are controlled by the manner in which motion profiles are acquired
and the extent to which such knowledge is available across an ad hoc network.
Coordination Models Applicable to Ad Hoc Networks
While coordination models represent a wide research area, I focus on
models that have a practical application to ad hoc networks. Ad hoc networks
make providing strong guarantees a difficult task due to their dynamic
character. The core of the consistency guarantees in my previous work was
provided by a Linda-like coordination model adapted to ad hoc networking. My
intentions are to investigate such coordination mechanisms further and develop
new models to define semantics of interactions in ad hoc networks. I am
particularly interested in investigating distributed inter-process communication
mechanisms and studying their semantics in mobile networks.
The three research efforts are integrated in the SPAWN middleware. The middleware
is under continuous development, as reserch findings are added to the paltform at
all three layers. Please check the SPAWN middleware webpage
here.
Context Aware Cell Phones
Well, this is no longer at the university so I will only say that I am trying to
use my cell phone to replace all my keys, credit cards, remote controls, and much
more. Kinf of anytime anywhere computing by enabling the cell phone to help me
right here and right now.
Information Rich Digital Television
These are only some wild ideas that need a lot more work before they even become
a plan. The question is: what ifwhen you watch a movie or a live broadcast, you'd
like to get some more info about James Bond's watch, the car a guy is parking in the
background image, next to the Eiffel Tower, or Mark Bulger's equipment on a Monday
night football game. In a digital TV era you'd pause the image and click on the
object of interest...