gossip - Smart Gossip: An Adaptive Gossip-based...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon
Smart Gossip: An Adaptive Gossip-based Broadcasting Service for Sensor Networks Pradeep Kyasanur Romit Roy Choudhury Indranil Gupta Department of Computer Science, University of Illinois at Urbana-Champaign Email: { kyasanur,croy } @uiuc.edu, indy@cs.uiuc.edu Abstract — A network-wide broadcast service is often used for information dissemination in sensor networks. Sensor networks are typically energy-constrained and prone to failures. In view of these constraints, the broadcast service should minimize energy consumption by reducing redundant transmissions, and be tolerant to frequent node and link failures. We propose “Smart Gossip”, a probabilistic protocol that offers a broadcast service with low overheads. Smart gossip automatically and dynamically adapts transmission probabilities based on the underlying network topology. The protocol is capable of coping with wireless losses and unpredictable node failures that affect network connectivity over time. The resulting protocol is completely decentralized. We present thorough experimental results to evaluate our “Smart Gossip” proposal, and demonstrate its benefits over existing protocols. I. INTRODUCTION Several sensor network applications rely heavily on a network-wide broadcast service for disseminating in- formation. For example, sink nodes may often need to broadcast code-updates, TAG-type queries, alarms, etc., to sensor nodes. Flooding, considered as the simplest means of broadcasting, does not prove to be applicable in the context of sensor networks. This is because flooding leads to collisions and redundant packet receptions [1] that together deplete sensors of valuable battery power. An ideal solution for sensor networks needs to be one that delivers a single copy of the broadcast packet to each sensor, using minimum number of transmissions. There have been several approaches that aspire toward this goal, broadly classified into deterministic and prob- abilistic approaches. Several deterministic approaches have aimed to mini- mize redundant communications by identifying a suitable subset of nodes, and assigning them the responsibility This research was supported in part by Vodafone and Motorola graduate fellowships, and by NSF grants ITR CMS-0427089 and CAREER CNS-04448246. of forwarding messages. The deterministic subset of nodes may either be cluster-heads in the network graph [2], multi-point relays [3], or members of a backbone tree [4] – whatever the name, they essentially build a connected dominating set of the graph, D G , and use it to disseminate information. However, this class of solutions suffer from two main problems. First, if a member of D G fails, sensors that depend on the failed member may not receive the broadcast message. This affects the reliability of the network, and requires expensive mechanisms to maintain D G over time. Second, using only members of D G to disseminate information affects the load-balancing properties of the network, and drains the battery at some sensors much earlier than others. Sensor network appli-
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 2
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 10

gossip - Smart Gossip: An Adaptive Gossip-based...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online