|
|
|
|
|
|
OPNET is a registered |
Comparison of
Optical Switching Technologies Optical Packet Switching (OPS) and Optical Burst
Switching (OBS) technologies make more efficient use of bandwidth as compared
to Optical Circuit Switching (OCS). OPNET will be used to model OPS, OBS and
OCS network architectures. We will compare the performance of these networks
using combinations of traffic distributions and protocols. This research will
help in identifying the conditions for optimum performance of each network. On-Demand QoS
Path Framework We proposed a new QoS framework, called the On-Demand QoS Path framework (ODP). It provides end-to-end QoS guarantees to individual flows with minimal overhead, while keeping the scalability characteristic of DiffServ. ODP exercises per-flow admission control and end-to-end bandwidth reservation at the edge of a network and only differentiates service types in the core of the network. In addition, to adapt to dynamically changing traffic load, ODP monitors the bandwidth utilization of a network and performs dynamic bandwidth reconfiguration in the network core based on the monitored bandwidth utilization. Through simulations, the performance of ODP is investigated and compared with that of IntServ with RSVP, DiffServ, DiffServ with Bandwidth Broker, IntServ with Aggregated RSVP frameworks. The simulation results showed that ODP provides end-to-end QoS guarantees to individual flows with small overhead. We are using OPNET to run more extensive simulations. Multi-viewing-point
Target Tracking in Sensor Networks We proposed and investigated a protocol that
selectively activate a group of sensors
to form a monitoring structure around a mobile target such that activated sensors monitor
the target from multiple viewing
angles. This protocol assumes that a large number of sensors with varying
capabilities (e.g., different sensing
and communication capabilities, and different mobility capabilities) are
randomly deployed, creating a large-scale and
dynamic sensor
network environment. It also assumes that sensors
can take one of
the three different modes (i.e. the
sleeping mode, the listening mode, and the active mode). In this protocol, upon detecting a mobile target, active
sensors broadcast notifications to their neighboring sensors
and activate neighboring sensors in the
listening mode in order to form a monitoring structure.
Based on the information contained in the
notifications (e.g., viewing angles to the target, distance to the target
etc.), each activated sensor autonomously decides whether to stay in the
active mode or switch to the sleeping mode. This distributed and autonomous activation
process results in the
self-organization of a monitoring structure that
cannot be achieved by any individual sensor. We have considered different scenarios based
on different sensor capabilities and developed a set of distributed protocols to selectively activate sensors and form an effective monitoring structure. We are going to build simulation models using
OPNET. Self-organizing
Target Guarding using Mobile Sensors We proposed and investigated a protocol to monitor and protect a mobile target in a large-scale and dynamic sensor network consisting of a large number of mobile
sensors with varying capabilities. This protocol enables mobile
sensors to self-organize, without the central controller, into a largest possible perimeter
that surrounds a mobile target (or mobile targets) such that no intruder can penetrate the perimeter without
being detected. In designing the proposed
protocol, mobile sensors and mobile targets are modeled as particles that exist in the nature (such as molecular),
and the repelling and attracting forces that exist among the particles in the
nature (such as inter-molecular force) are introduced to determine the
direction of the movement of sensors to form a perimeter. By applying inter-molecular force on
mobile sensors, mobile sensors
form a perimeter autonomously in a distributed
manner. We have designed the protocol that controls the movement
of mobile sensors to form an optimal perimeter around mobile targets.
We are going to use OPNET to build simulation models. Coverage-Aware
Self-Coordination in Sensor Networks We proposed and investigated a new protocol, Coverage-Aware Self-Coordination (CASC) protocol, to efficiently
provide enough network surveillance of the area of interests. The CASC
protocol achieves the goal by selectively activating sensors according to
specific requirements of applications. In order to select which sensors
should be activated, each
sensor evaluates its own contribution to network
surveillance in a
distributed manner. Based on its contribution to network
surveillance, each sensor determines if it
should stay active and continue monitoring or if it should deactivate itself. Sensors with higher contribution to
network surveillance stay active with higher probability, thus fewer active sensors are needed to
guarantee enough surveillance. We are currently conducting preliminary simulations to evaluate
the performance of the proposed CASC protocol. Simulation results show that the CASC protocol guarantees enough network surveillance by
activating a small number of sensors. More extensive simulations in OPNET will be performed to investigate the performance of the CASC protocol under dynamic environments. Data
Dissemination in Sensor Networks We proposed and investigated a new protocol, Rendezvous Track protocol, for sensors to determine where
to store sensing data and for users to determine where to retrieve the sensing
data. In the proposed protocol, each
sensor applies
a hash function to the attributes of the sensing data, yielding the location
of a group of sensors to store the sensing data. When retrieving sensing data,
a user applies the
same hash function to the attributes of the desired sensing data to obtain
the location of a group of sensors that store the desired sensing data. This
group of sensors serve as rendezvous sensors for sensing data with some
particular attributes so that sensing data and user queries containing the
same attributes will always converge at the same set of rendezvous sensors,
leading to successful sensing data retrieval. We also developed a mechanism
through which the proposed protocol dynamically adjusts the group of sensors
to store sensing data based on different
user query patterns. We are currently designing and
implementing OPNET simulation models to investigate the scalability, robustness,
adaptability, and efficiency of the proposed protocol. |
