Sensor Networks Storage Sanket Totala Sudarshan Jagannathan
Outline Introduction Storage Mechanisms Multi-Resolution Storage Two-Tier Storage Conclusion
Introduction Sensor Nodes Low Memory Low Power Low processing capability Inapproachability Sensor Networks Efficiency requirements Highly dense
Introduction Sensor Storage Large number of events Query handling capability required Streaming data Aging mechanism Data organization
Sensor Network Storage - Mechanism Centralized Storage Central server Ample power Sufficient storage Single point of failure Fast Query processing High communication Sparse Networks Infrequent events Low Data Transfer Scalability problems
Sensor Network Storage - Mechanism Distributed Storage Storage at each node Local computation Scalable Dense Networks Frequent events High Data Transfer Slower Query processing Flooding Distributed Indexing Drill-down querying
Multi-Resolution Storage Multi-Resolution summarization Construct data summaries Hierarchy construction Drill-down query evaluation Narrowing search space Spatio-temporal compression Data aging Efficient storage utilization Data Degradation
Multi-Resolution summarization Temporal Summarization Exploits Temporal redundancy Computation overhead No communication overhead Spatial Summarization Exploits Spatial redundancy Data Summarization every level Hierarchy construction
Drill-down query evaluation Spatial compression Finer data view with every level. Reduced search cost Query routing Sub region selection More accurate result
Data Aging Long deployment Limited storage Efficient resource utilization Fast query processing Accuracy Higher level, Higher time periods
Data Aging – Algorithms Omniscient Algorithm Data sets available Full global knowledge required Query specific Training Algorithm Data sets available Data set partitioned Training data (available during sensor deployment) Test data Greedy Algorithm Data set unavailable Assigns weights to data summaries
Outline – Two Tier Sensor Storage Design Considerations and Principles System Design Architecture Data structures Data Storage Sensor Network Data Summarization Conclusion
Design Consideration and Principles The Three-Tier Model Bottom tier - Untethered sensor nodes Middle tier - Tethered sensor proxies Upper tier - Applications and user terminals
Design Consideration and Principles Principles Store locally, access globally Distinguish data from metadata Provide data-centric query support
Data Structures Used Skip Graph Ordered index In-place indexing Log n height Probabilistic balance Redundancy and resiliency
Data Structures Used Interval Skip Graph Extends skip graphs to store intervals Allows efficient searches Complexity of search is O(log n) Insertion cost of O(n) Sparse Interval Skip Graph
Data Storage Local Storage at Sensors Archival Stores: Collection of records Interval skip graphs used Efficient routing and query handling Operations: Create, Read & Delete
Sensor Data Summarization Data summaries – bind the storage at the remote sensor and the index at the proxy Each update from a sensor to the proxy includes The summary Time period corresponding to the summary The start and end offsets for the flash archive
Sensor Summarization Adaptive Summarization Balances the cost of sending updates against the cost of false positives Summarization Parameters The interval over which summaries of the data are constructed and transmitted to the proxy The size of the application specific summary.
Conclusion Multi-tier nature of sensor networks Large amounts of events and data Lossy Approach Two-Tiered Approach Low overheads Decentralized and Hierarchical storage Possible Solution Combine lossy nature of Multi-resolution with routing techniques of Two-tiered approach.