Cushion: autonomically adaptive data fusion in wireless sensor networks
Jungmin So, Jintae Kim, Indranil Sen Gupta · 2005
In a typical in-network aggregation problem, data originates from multiple source nodes, and moves towards single sink node or root node. Along the way, i.e., inside the network, the data may be partially aggregated, thus reducing message overhead. Two well-known classes of solutions to this problem are: tree-based aggregation and multipath aggregation. The autonomically adaptive data fusion solution, called Cushion, contains two protocols that span a continuous spectrum between these two design points and can further increase reliability over multi-path approach using redundant transmissions. The design is motivated of Cushion by first discussing the complementary advantages and disadvantages of the two design points.