Delta Compression in Wireless Sensor Networks

Wireless sensor networks and other data stream-based systems may be subject to higher input rates than their resources can handle. When overloaded, the system must utilize some scheme to overcome this difficulty while maintaining quality, throughput, and latency requirements. One common way of dealing with this problem is the use of compression to reduce the amount of data flowing through the system. Due to the limited storage and processing power of wireless sensors, the methods for compression within such networks must be have low associated costs for performing the compression. Also to achieve savings within the network compression must be done without global knowledge of the data. Many current methods sacrifice latency [1], scalability [2], or quality [3] to achieve improvements in bandwidth utilization and throughput. We propose a two-layer compression scheme that reduces the amount of data flowing through the network while avoiding the above sacrifices. In the first layer, the data is transformed using delta compression then the second layer compresses the delta values. No delay is needed to collect similar values and the data can be decompressed at the sinks or base stations without any loss in granularity or accuracy. Also the compression can be done on any size data set with any range of possible values. Preliminary simulations show promising results for several real-life data sets.