Collaborative Storage Management In Sensor Networks

Sameer Tilak, Nael B. Abu-Ghazaleh, Wendi Rabiner Heinzelman

Abstract

In this paper, we consider a class of sensor networks where the data is not required in real-time by an observer; for example, a sensor network monitoring a scientific phenomenon for later play back and analysis. In such networks, the data must be stored in the network. Thus, in addition to battery power, storage is a primary resource: the useful lifetime of the network is constrained by its ability to store the generated data samples. We explore the use of collaborative storage technique to efficiently manage data in storage constrained sensor networks. The proposed collaborative storage technique takes advantage of spatial correlation among the data collected by nearby sensors to significantly reduce the size of the data near the data sources. We show that the proposed approach provides significant savings in the size of the stored data vs. local buffering, allowing the network to run for a longer time without running out of storage space and reducing the amount of data that will eventually be relayed to the observer. In addition, collaborative storage performs load balancing of the available storage space if data generation rates are not uniform across sensors (as would be the case in an event driven sensor network), or if the available storage varies across the network.

Introduction

Traditional research in Wireless Sensor Networks (WSNs) has primarily focused on energy conservation as the ultimate constraint. However, for a specific class of "delay-tolerant" applications—such as wildlife tracking (e.g., ZebraNet) or remote ecological monitoring—data is collected and stored locally until an observer can retrieve it. In these scenarios, storage capacity becomes a critical bottleneck. Sensors possess limited persistent memory; once full, they must either cease sensing or overwrite existing data, leading to information loss. Furthermore, spatial correlation in sensor data often leads to redundant storage, while event-driven data generation can cause localized storage "hotspots" that deplete rapidly while other nodes remain underutilized.

Proposed Storage Management Strategies

To address these challenges, we evaluate several storage management strategies:

  • Local Storage (LS): The baseline approach where each node buffers its own data. While it consumes no communication energy for storage, it is inefficient due to data redundancy.
  • Cluster-Based Collaborative Storage (CBCS): Nodes form clusters where members transmit data to a Cluster Head (CH). The CH exploits spatial correlation to aggregate and compress data before storage, reducing the overall footprint and distributing the storage burden.
  • Coordinated Approaches (CLS/CCS): These variants introduce a feedback loop where the CH dynamically adjusts member sampling rates based on observed redundancy.
  • Geographical Zoning: To maintain efficiency, the sensor field is divided into zones to ensure clusters remain localized, preventing high multi-hop communication costs.

Experimental Methodology

The proposed methods were evaluated using the NS-2 (Network Simulator 2) environment. Simulations were conducted in a 350m x 350m field with varying sensor densities (50, 100, and 150 nodes). Two primary scenarios were tested:

  1. Even Data Generation: All nodes sample at a uniform rate.
  2. Uneven Data Generation: Simulates localized events where specific nodes sample at twice the rate of others, testing the load-balancing capabilities of the protocols. Parameters were modeled after real-world projects like ZebraNet to ensure practical relevance.

Performance Evaluation

Storage Efficiency

The primary advantage of collaborative storage is the reduction of the data footprint. As shown in the figure below, the mean storage space consumed per sensor is significantly lower for CBCS and its coordinated variants (CLS, CCS) compared to the Local-Buffer baseline. In dense networks, these collaborative methods achieved approximately a 50% reduction in storage consumption, directly reflecting the aggregation ratios employed.

Mean storage space consumed across different network sizes and management strategies.

Network Lifetime and Load Balancing

The impact of storage management on network longevity is most evident in uneven data generation scenarios. Local storage leads to rapid depletion of "hotspot" nodes. In contrast, collaborative storage balances the data load across the cluster's collective memory. The following figure illustrates the percentage of sensors that have depleted their storage over time. Collaborative approaches significantly delay the onset of storage depletion, allowing the network to maintain 100% coverage for a much longer duration than local buffering.

Percentage of storage-depleted sensors over time, comparing local and collaborative approaches under even and uneven data generation.

Data Retrieval and Energy Trade-offs

Collaborative storage involves a trade-off: energy is expended early for data transmission to the CH, but this investment pays off during the data retrieval phase. Because the stored data is aggregated and smaller, the "Mean Collection Time" for an observer to pull data from the network is reduced by 40-50% in dense configurations. This reduction in data volume also translates to lower energy consumption during the final reach-back phase.

Conclusion and Future Work

This study demonstrates that in store-and-forward sensor networks, storage is a primary resource that must be managed as carefully as energy. Cluster-Based Collaborative Storage (CBCS) effectively mitigates storage bottlenecks by exploiting spatial correlation and performing load balancing. While these results are promising, future work should involve real-world implementation on hardware like Berkeley Motes to validate power profiles. Additionally, exploring "semantic" storage management—where data importance dictates compression levels—could further optimize resource utilization.

Summary updated

Abstract

In this paper, we consider a class of sensor networks where the data is not required in real-time by an observer; for example, a sensor network monitoring a scientific phenomenon for later play back and analysis. In such networks, the data must be stored in the network. Thus, in addition to battery power, storage is a primary resource: the useful lifetime of the network is constrained by its ability to store the generated data samples. We explore the use of collaborative storage technique to efficiently manage data in storage constrained sensor networks. The proposed collaborative storage technique takes advantage of spatial correlation among the data collected by nearby sensors to significantly reduce the size of the data near the data sources. We show that the proposed approach provides significant savings in the size of the stored data vs. local buffering, allowing the network to run for a longer time without running out of storage space and reducing the amount of data that will eventually be relayed to the observer. In addition, collaborative storage performs load balancing of the available storage space if data generation rates are not uniform across sensors (as would be the case in an event driven sensor network), or if the available storage varies across the network.
BibTeX
@article{tilak2005collaborative,
  author = {Sameer Tilak and Nael B. Abu-Ghazaleh and Wendi Rabiner Heinzelman},
  title = {{Collaborative Storage Management In Sensor Networks}},
  journal = {Int. J. Ad Hoc Ubiquitous Comput.},
  year = {2005},
  url = {https://openreview.net/forum?id=5RyZv_FwYqPT},
}
Figure