ViSE team is pictured from left to right: Navin Sharma (graduate student), Prashant Shenoy (PI), David Irwin (Research Scientist), and Michael Zink (Co-PI). Deepak Ganesan and Jim Kurose (both Co-PIs) are not pictured.

The UMass GENI team is at the forefront of a trend.  A few years from now, widespread deployment of programmable sensors may allow us to live within an environment immensely rich in realtime environmental information, ranging from fine-grain, local weather information that improves crop yields right down to sensors that continuously monitor our health and roll it up into live, world-wide maps of wellness and disease.

As Principal Investigator, Prashant Shenoy is leading the UMass-Amherst team incorporating their Virtual Sensing Environment (ViSE) into GENI prototyping.

The ViSE team has three ambitious goals: create sensor virtualization technology for GENI, virtualize sensors to enable sharing and flexible experimentation, and integrate their prototypes into GENI so researchers can control a suite of sensors as part of end-to-end GENI experiments. To do so, they rely on ORCA, a prototype GENI control framework being created by Jeff Chase of Duke University.

“Our research group has long focused on prototyping real systems and understanding their behavior using analysis and experimentation,” Shenoy says. That focus on building real systems has helped ViSE make rapid progress since beginning their GENI project in October, 2008. The ViSE team has adapted prototype sensor nodes built by students at the Center for Collaborative Adaptive Sensing of the Atmosphere (CASA) to support virtualization and the “deeply programmable” nature of GENI.

In collaboration with CASA, the team has deployed these nodes on the UMass-Amherst Computer Science building, the Mount Toby fire tower, and CASA’s MA1 tower, and connected them with long-distance WiFi using directional antennas.  Each node includes a portable weather sensing radar, a DavisPro weather station, and a pan-tilt-zoom camera.

The ViSE team is integrating their prototypes into GENI for broader use by researchers, who will be able to use it to experiment with their own ideas for improving weather monitoring. The combination of resources will allow a broad range of researchers to experiment with devices that provide both sensing and actuation (control), giving them a range of data rates from low-duty-cycle weather stations to high data rate controllable radars connected to powerful computing capabilities and communication. Users will be able test their own software by downloading it into testbed nodes and directing the node’s sensing and communication of data. Because the sensor network is virtualized, multiple users will be able to share the resources concurrently and efficiently.

The team expects to continue their rapid progress given their considerable experience in virtualizing computing platforms and federating virtualized networks combined with their early success in constructing, deploying, and networking three nodes.  This summer, researchers across the United States  will be able to log in and request slices on the ViSE testbed to experiment with and test sensor-based services. If successful, ViSE prototypes may be expanded to wider-scale deployment for large-scale experimentation.

ViSE Node Deployment on UMass Amherst CS Building (with Mt. Toby in the background)