Usable and efficient abstractions for wireless sensor networks

点击翻译按钮获取中文摘要

Wireless sensor networks are formed by inexpensive, underpowered, and rather unreliable wireless devices/nodes that are deployed in all kinds of environments to allow the remote acquisition of sensed data. Transmissions are expensive because they require a significant amount of the energy available to the nodes. To cope with this problem, sensed data, such as temperature, motion, pressure, etc. are collected and aggregated within the network, with the intention to reduce the total amount of data that needs to be eventually transmitted. Application programmers who wish to develop applications for wireless sensors networks need to be able to translate the application requirements into computation and communication tasks for each node. Yet, because of the limited resources of those nodes, they do not have the luxuries afforded by high-end computing devices that run fully capable operating systems. The question thus becomes what should be the minimum set of capabilities that would allow a wide range of applications to be developed, and are possible to incorporate in the limited resources of sensor nodes. At the same time, one should be able to properly and easily reason about the application logic. That is why the minimum set of capabilities should be accompanied by a model of computation and communication. In essence, the communication and computation model acts as the framework in which applications will be developed. Our proposal addresses precisely this issue, i.e., what abstractions and model for development of applications is suitable for wireless sensor nodes. We use our experience that spans from theoretical results to implementations of such networks. The proposal places emphasis on how to leverage the large number of sensor nodes to provide a "communal" form of resources over which an application can be developed, while at the same time being capable to deal with failures and energy saving strategies in a fashion that does not obstruct the applications.