This webpage documents my presentation of
"Matching Data Dissemination Algorithms to Application Requirements" by John Heidemann, Fabio Silva, and
Deborah Estrin (SenSys 2003): Topics in Time-sensitive Distributed
Systems, taught by Charles 'Buck' Krasic.
The paper’s main goal
is to address the problem of how to choose an appropriate algorithm for a
sensor networks application. Since there are an increasing number of
applications for sensor networks and also the collateral effects in choosing
effective algorithms for these applications, it is important for an application
developer to have a guideline on which algorithm is suitable for his application
(assuming that he has knowledge about the available classes of algorithm for
sensor networks out there).
•
Data
dissemination approaches in sensor networks have adopted application-specific,
data-centric communications protocols in order to reduce overhead by
avoiding levels of abstraction and to support application involvement in
communication. In addition, application-specific constraints and optimizations
greatly reduce communications cost by replacing communication with computation
in the network. However, as number of protocols and sophistication of
applications grows, choice of communication algorithms becomes a problem.
•
Sensor network
applications have different needs such as different traffic patterns
(many-to-one, many-to-many, one-to-many, one-to-one); different data rates
(fixed and variable, frequent and infrequent). Furthermore, applications must
be robust to change, e.g. wireless links come and go, nodes fail or move. The
solution is to address
application-specific requirements by matching routing algorithms to application
requirements.
•
Diffusion Routing
Algorithms have some typical characteristics including localized algorithms,
named data, and support for in-network processing, based on set of abstract
•
Multiple
Diffusion Routing Algorithms: Two-Phase Pull Diffusion, One Phase Pull
Diffusion, Push Diffusion, GEAR.
•
Systematic
Evaluation provides a methodology to evaluate a range of algorithms based
simulation and emulation.
The slides for my
presentation (in .ppt format) can be found here.
Questions about the paper:
·
The research
question put forth by the paper “How can communication be robust
but also efficient for many different applications? ” though
valid and helpful, however seems contradictory the goal of sensor networks that
is application specific.
·
While the general
methodology for systematic evaluation brings some valid results reflected on
the obvious effects of choosing algorithm correspond to applications, there are
much more room for improvement for the application developers to have a more
generic means to answer the question whether alg. A is best for app. B or not?
Or what kind of algo. can best suit app. B?….
·
The number of sensor
networks applications may be unbounded, then is there a general methodology to match class(es) of algorithm to applications
?