This webpage documents my presentation of
"Next Century Challenges: Scalable Coordination in Sensor Networks" by Deborah Estrin, Ramesh Govindan, John
Heidemann and Satish Kumar Mobicomm 1999 for
CPSC 538A: Topics in Time-sensitive
Distributed Systems, taught by Charles 'Buck' Krasic.
The paper’s main goal is
to serve as an introductory material for sensor networks, its visions, its
applications and also its challenges. It also introduced the basic algorithms
and information structures for sensor networks applications. Overall, it helps improve
the general public understanding about sensor networks and serves as a
background for further research in the field.
The story begins like this: Integrated low-power sensing
devices will permit remote object monitoring and tracking in many different
contexts: in the field (vehicles, equipment, personnel), the office building (projectors,
furniture, books, people), the hospital ward (syringes, bandages, IVs) and the
factory floor (motors, small robotic devices). Networking these sensors empowering
them with the ability to coordinate amongst them on a larger sensing task will
revolutionize information gathering and processing in many situations. Large
scale, dynamically changing, and robust sensor colonies can be deployed in hospitable
physical environments such as remote geographic regions or toxic urban
locations; even in large industrial plants, aircraft interiors etc.
Several aspects of this scenario present systems
design challenges different from those posed by existing computer networks such
as:
·
The sheer
numbers of these devices, and their unattended deployment, will
preclude reliance on broadcast communication or the configuration currently
needed to deploy and operate networked devices.
·
Devices may be
battery constrained or subject to hostile environments, so individual device
failure will be a regular or common event.
·
The configuration
devices will frequently change in terms of position, reachability, power availability,
and even task details.
·
Finally, because
these devices interact with the physical environment, they, and the network as
a whole, will experience a significant range of task dynamics.
Many of the lessons learned from Internet and mobile network
design will be applicable to designing sensor network applications. However, sensor
networks have different enough requirements to at least warrant re-considering
the overall structure of applications and services. Specifically, we believe
there are significant robustness and scalability advantages to designing applications
using localized algorithms where sensors only interact with other
sensors in restricted vicinity, but nevertheless collectively achieve a desired
global objective. Directed diffusion is introduced as a promising information model
for describing localized algorithms.
The slides for my
presentation (in .ppt format) can be found here.
Questions about the paper:
·
Buck remarked
that Sensor network today likes the Internet 30 years ago! How it is going to
evolve?
·
Sensor devices
are easy to fail, leave problems for applications, network management (e.g. how
to exclude those failed sensor, how to distinguish between the failed nodes or
the nodes temporarily hidden by obstacles …), and sensor devices also may
pollute the environment.
·
Since sensor
networks is data-centric and application-specific, then when the
number of applications increase, the pressure on the research community to
create more specific algorithms for meet the requirements of applications also increase.
Then how to map algorithms to applications efficiently?
·
In an environment
where the failure rate of sensors is very high (or nodes may reduce its energy
level very quickly), would localized algorithms cause instability to the sensor
networks? And how can the tradeoff between stability and fidelity be achieved satisfactorily?
·
Is it sufficient
to design sensor network applications using Internet technologies coupled with
ad-hoc routing mechanisms?
Discussion of related works: