An Overview of Directed Diffusion

Traditional sensing models assume one or a few powerful sensors and centralized computation. Today, technological trends enable the creation of inexpensive, small, intelligent devices for sensing and actuation. If many small sensors can work together as a sensor network, they provide several advantages over traditional centralized sensing. By placing the sensor close to the object being sensed, signal processing and target discrimination problems in sensing can be greatly simplified. By communicating over several short hops rather than one long hop, energy consumed in communication can be reduced [29]. Moreover, by processing data in the network, often the amount of data transfered can be reduced, further saving energy [22]. Motivated by robustness, scaling, and energy efficiency requirements, this paper examines a new data dissemination paradigm for such sensor networks. This paradigm, which we call directed diffusion1, is data-centric. Data generated by sensor nodes is named by attribute-value pairs. A node requests data by sending interests for named data. Data matching the interest is then “drawn” down towards that node. Intermediate nodes can cache, or transform data, and may direct interests based on previously cached data (Section 3). Directed diffusion is significantly different from IP-style communication where nodes are identified by their endpoints, and inter-node communication is layered on an end-