Tenet: A Sensor Network Architecture


Many sensor networks consist of two classes of nodes. Small form-factor and resource constrained wireless sensor nodes are deployed close to the phenomena being observed. Larger PC-class devices form the processing and communication backbone for the network. Tenet is a formalization of this two-tier architecture and comes with a general-purpose software stack that has been used across different types of sensor network applications.

Programming Tenet

A linear dataflow language is used to program the sensor nodes in Tenet. A library of reusable and composable software modules is provided covering a varity of tasks such as sensing, processing, filtering, and thresholding the data. The language does not allow nodes to aggregate data from multiple nodes. This means the nodes perform simple in-node processing of data as specified by the data flow program and send the data to the base station. All the aggregation on data from across the nodes is done at the base station. We have found that applications written in this centralized architecture can be as efficient as distributed applications, even for complex applications like track tracking.

Duty-Cycling Tenet

Keeping the radio on when no communication is taking place can be a major source of energy drain in a sensor node. Turning the radio off during those idle times, or radio duty-cycling, can result in large energy savings in sensor nodes and hence significantly longer network lifetime. We designed Application Informed Energy Management (AEM), a radio-dutycycling mechanism that exploits the properties of Tenet and meets Tenet's requirements. Tenet's simple dataflow programming language means we can easily analyze the application and identify its data generation pattern. AEM statically analyzes and infers the traffic profile for the application and accordingly tunes the duty-cycling protocol. AEM uses elastic transmission and reception schedules that allows it to adapt to dynamics while enabling bounded latency of event detection. Our experiments show that AEM achieves 1-3% duty-cycles, while satisfying these three hard requirements: dynamic multi-hop routing and tasking, multiple concurrent applications, nad reliable end-to-end delivery.


Jeongyeup Paek, Ben Greenstein, Omprakash Gnawali, Ki-Young Jang, August Joki, Marcos Vieira, John Hicks, Deborah Estrin, Ramesh Govindan, and Eddie Kohler, The Tenet Architecture for Tiered Sensor Networks, ACM Transactions on Sensor Networks (TOSN), Vol. 6, No. 4, 2010.
Omprakash Gnawali, Jongkeun Na, and Ramesh Govindan, Application-Informed Radio Duty-Cycling in a Re-Taskable Multi-User Sensing System, In Proceedings of the ACM/IEEE International Conference on Information Processing in Sensor Networks (IPSN 2009), San Francisco, CA, April 13-16, 2009. Acceptance Rate - 21/117 (17.9%)
Omprakash Gnawali, Ben Greenstein, Ki-Young Jang, August Joki, Jeongyeup Paek, Marcos Vieira, Deborah Estrin, Ramesh Govindan, and Eddie Kohler, The TENET Architecture for Tiered Sensor Networks, In Proceedings of the ACM Conference on Embedded Networked Sensor Systems (SenSys 2006), Boulder, CO, October 31 - November 3, 2006. Acceptance Rate - 24/122 (19.7%)
Jeongyeup Paek, Omprakash Gnawali, Ki-Young Jang, Daniel Nishimura, Ramesh Govindan, John Caffrey, Mazen Wahbeh, and Sami Masri, A Programmable Wireless Sensing System for Structural Monitoring, In Proceedings of the Fourth World Conference on Structural Control and Monitoring (4WCSCM), San Diego, CA, July 11-13, 2006.

Last updated: November 11, 2010