NeTS: Small: Realizing an Architecture for Community Sensor Grids

The proposed work involving two undergraduate students in the development of the community sensor grid architecture is aligned with the original objective of the project it is meant to supplement, i.e., to create core capabilities and expertise is Distributed Sensor Information Technology and Embbeded Systems. The students will work with the PI, co-PI and Graduate Resaerch Assistants in the Operating Systems Laboratory under the ASU School of Computing, Informatics, and Decision Systems Engineering. Both undergraduate students are will be assigned work for 20 hours per week during the Summer 2012 and Fall 2012. One student who has been working with a Graduate Research Assistant on the laboratory experiments involving on-demand formation of wireless PAN underlay, will continue to assist in the development process of the enabling middleware for mote-level virtualization, especially with an emphasis on collection tree formation and modular thread binary or byte code dissemination. The other student is expected to replace a current REU student, will continue on his code base for map-based discovery of community sensing services and formation of the peer-topeer control network and work with another Graduate Research Assistant on safe thread binary code synthesis and run-time exception handling support.

Wireless sensor networks have been traditionally designed to be privately owned and used. Hence the two hallmark features of sensor networks, namely custom designed network applications with optimized power-performance trade-offs and the collaborative in-network processing for sensor data fusion and storage, are not achievable beyond the boundary of the users' administrative domain. While there has been a recent thrust on Internet-integration of low-power personal area networks such as the sensornets, the notion of sensor grid is at a nascent stage. If a community accessible sensor grid across multiple administrative domains is desired, the current state-of-the-art is to interconnect several isolated sensornet clouds with Internet gatewaysusing published services as building blocks for data-only accesses. In this proposal, we advocate a different notion of sensor grid that preserves the operational continuity of the participating sensornets and yet allows controlled sharing of mote hardware capabilities spanning multiple administrative domains. This calls for the creation of a transient sensornet underlay for stitching together virtualized nodes donated by participating sensor subnets. Possible approaches for mote-level virtualization support are discussed for several existing sensornet OS platforms. We also propose the formation of a P2P overlay network of the participating Internet-sensornet gateways to enable the formation, operation and dissolution of the transient underlay. Intellectual Merit: The intellectual merit of this proposal primarily lies in the new notion of programmability and resource utilization over an individually owned but community shared sensor grid. This is expected to be a leap forward towards the Internet-integration of wireless embedded sensornets. In addition, the proposed research introduces the capability of virtualization and slicing into the impoverished sensor nodes, also known as motes. This may be a step towards extending GENI compliance to low-power personal area networks. It will also impact the development of open source software that supports the P2P sensor grid architecture, as well as the possibility of standardization effort. Initial activities under this project also include a comparison of capabilities and performance of the existing and forthcoming embedded network operating systems and middleware platforms to bring forth their advantages and deficiencies, which will hopefully impact their future development. Broader Impact: In terms of pushing technology to mass market, the P2P structure enables democratization and allows underpowered citizens to participate and harness the power of global sensing. Resulting commercialization and mass marketing of embedded sensor devices will serve the dual role of securing property as well as participating in community sensor grids. The technology is expected to contribute to urban sensing, emergency response system, and homeland security. The permeation of the technology can impact society and underrepresented groups. The impact of making global sensing infrastructure to ordinary citizens for business or pleasure would favor future technological applications. Impact on Education: The results from this research will be used in courses at the undergraduate and graduate level such as networking, operating systems, security, and embedded systems. Developing such courses will be enabled from the experience gained from the proposed research. Such courses will be valuable in training personnel for future embedded applications. The core capabilities in embedded systems and networked sensing growing out of this project will strengthen the education of ASU undergraduate and graduate students. The PI and co-PI will actively work with More Graduate Education at Mountain States Alliance (MGE@MSA) to promote participation of minority students in the proposed research and pursue REU to recruit and support student from min