Matching Data Dissemination Algorithms to Application Requirements John Heidermann, Fabio Silva, Deborah Estrin Presented By: Bryan Wong

Outline Introduction Problem Description Diffusion Routing Algorithms Evaluation Conclusion

Introduction Data dissemination algorithms are application specific Reduces communications costs by replacing communication with computation in the network As number of protocols and sophistication of applications grows, choice of communication algorithm becomes a problem

Problem Description How can diffusion address application- specific requirements?

Robustness Requirements Applications must be robust to change: Wireless links come and go Nodes fail or move How can communication be robust but also efficient for many different applications?

Application Requirements Sensor network applications have different needs Different traffic patterns (many-to-one, many-to- many, one-to-many, one-to-one) Different data rates (fixed and variable, frequent and infrequent)

Solution Match routing algorithms to application requirements

Multiple Diffusion Routing Algorithms Two-Phase Pull Diffusion One Phase Pull Diffusion Push Diffusion GEAR

Two-phase pull diffusion Initial diffusion implementation Periodically floods data sink’s interests and exploratory data

GEAR Adds support for geographically scoped queries If nodes know their locations, then geographic queries can influence data dissemination Replaces network wide communication with geographically constrained communication

Push Diffusion Reverses the roles in the publish/subscribe API Floods only exploratory data messages

One-phase pull diffusion Subscriber based system that avoids one of the two phases of flooding in two-phase pull Only floods interests No exploratory messages

Sample Applications Push reduces message count by ~60% compared to two phase pull

Sample Applications GEAR reduces message count by ~40%

Systematic Evaluation

One-phase pull is best with many sources, few sinks Push works best with many sinks and few sources

Conclusions The break even point between the two algorithms depends upon specific control message frequency as well as application data rates For networks with more than a few dozen nodes, the benefits of geographically-scoped queries can outweigh other algorithmic choices

References ers/Research%20Review/015_Silva.pdf