Scalable medium access control for in-network data aggregation

  • Authors:

  • Jamie Macbeth;Majid Sarrafzadeh

  • Affiliations:

  • University of California, Los Angeles, CA, USA;University of California, Los Angeles, CA, USA

  • Venue:
  • Proceedings of the fifth international workshop on Foundations of mobile computing
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

We present conflict-free and contention-based medium access control (MAC) protocols designed for resource-aware data collection in sensor networks. We are interested in the performance of these schemes when used in in-network data aggregation systems. We introduce a Listen-and-Suppress (LAS) MAC protocol paradigm which can conserve network and node resources and cut delays through the interaction between the constituent nodes. In LAS-TDMA and LAS-CSMA, nodes listen to the channel and suppress their transmissions and sleep if their data is not needed. Under these conditions, we compare conflict-free scheduling and random scheduling in a general setting along several performance metrics. We find that, for conflict-free scheduling, collecting the aggregate minimum or maximum of a data value in records residing on n nodes in the network requires, on average, O(lg n) record transmissions and O(n lg n) listens collectively. Without our scheme, n transmissions and n2 collective listens are required. We simulate in the random scheduling domain, and examine how delay can be reduced by increasing the offered load on the channel at the cost of greater power dissipation due to collisions. For networks of 20 nodes, LAS-CSMA reduces the average delay by 58% in comparison to CSMA, and for networks of 100 nodes, it reduces the average delay by 80%.