Scheduled System Maintenance:
Some services will be unavailable Sunday, March 29th through Monday, March 30th. We apologize for the inconvenience.
By Topic

Crash-Tolerant Collision-Free Data Aggregation Scheduling for Wireless Sensor Networks

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

The purchase and pricing options are temporarily unavailable. Please try again later.
1 Author(s)
Jhumka, A. ; Dept. of Comput. Sci., Univ. of Warwick, Coventry, UK

Data aggregation scheduling, or converge cast, is a fundamental pattern of communication in wireless sensor networks (WSNs), where sensor nodes aggregate and relay data to a sink node. For WSN applications that require fast response time, it is imperative that the data reaches the sink as fast as possible. For such timeliness guarantees, TDMA-based scheduling can be used to assign time slots to nodes in which they can transmit messages. However, any slot assignment approach needs to be cognisant of the fact that crash failures can occur (e.g., due to battery exhaustion, defective hardware). In this paper, we study the design of such data aggregation scheduling (converge cast) protocols. We make the following contributions: (i) we identify a necessary condition to solve the converge cast problem, (ii) we introduce two versions of the converge cast problem, namely (a) a strong version, and (b) a weak version , (iii) we show that the strong converge cast problem cannot be solved, (iv) we show that deterministic weak converge cast cannot be solved in presence of crash failures, (v) we show that there is no d-local algorithm that solves stabilising weak converge cast in presence of crash failures, (vi) we provide a modular d-local algorithm that solves stabilising weak converge cast in presence of crash failures where d is the network radius, and (vii) we show how specific instantiations of parameters can lead to an d-local algorithm that achieves more efficient stabilization. Our contributions are novel: (i) the first contribution (necessary condition) provides the theoretical basis which explains the structure of existing converge cast algorithms, and (ii) the study of converge cast in presence of crash failures has not previously been studied.

Published in:

Reliable Distributed Systems, 2010 29th IEEE Symposium on

Date of Conference:

Oct. 31 2010-Nov. 3 2010