What do 15,000 Java developers have in common? Besides a love of object-oriented programming languages and an oddly-shaped mascot named Duke, Java developers act en masse. How do we know? Sentilla was commissioned by Sun Microsystems to instrument the Moscone Center in San Francisco with over 200 pervasive computers during the 2008 JavaOne conference. Sun wanted to have greater insight into the behavior of attendees during the conference so that organizers could adjust energy-usage in real-time and better plan for future events. Each morning, John Gage presented Sentilla's findings. At the close of the conference, he exclaimed, “We have enough data to establish new patterns of behavior.” For instance, Gage and organizers from Sun were able to correlate the actions of the attendees to spikes in resource consumption, coming up with ways to save significant amounts of energy. Gage insists that all future conferences at Moscone -- whether from Sun, Apple, Microsoft, Oracle, or others -- should use Sentilla's solutions to minimize their environmental impact.

By knowing how and when people use a building's resources, energy usage can be tightly controlled. A new application brief from Sentilla discusses the impact of intelligent management, using resources exactly when and where they're needed rather than the wasteful practice of keeping everything on all the time.

Read the application brief >>

This entire application was built by Sentilla's engineers over the course of a couple of months using the Sentilla Software Suite and some custom hardware. We counted each person that entered or exited each room and monitored the environmental conditions including temperature, humidity, and light. Not only did we know when resources were in use, but we could also gauge the popularity of each session. Jonathan Schwartz, CEO of Sun, told EVP for Software, Rich Green, "That's going to be your performance review for the year. I want to make sure that everybody stays throughout the session, and you'll get a pass."

Dual infrared emitters and transmitters were mounted to each doorway in the Moscone Center (shown to the right). The devices analyzed the patterns by which each infrared beam was broken to determine the direction a person is walking and count each entrance and exit. The devices were mounted using magnets; the units snapped right onto the doorways in seconds and removed quickly at the end of the conference without any impact on the facility. Additional pervasive computers were placed on the stages and in the audience to analyze the humidity and temperature. With these sensors, the analysis showed each time that the H/VAC system turned on to cool the room, and the frequency of H/VAC events was correlated with the quantity of people in the room. Light sensors also showed us when the lights were left on (for example, in the evening), and even the start and stop times of each technical session.

The data was brought back to a webpage, and shown in each morning's keynote presentation. Below is the data from the first day's keynotes, where you can see the first keynote had strong attendance while the second keynote occurred while other sessions were in progress and thus had more flux.

Further data shows us the environmental conditions in a session room. Using the temperature, you can see the H/VAC cooling events. Using the light sensor data, you can see the changes in light as each session starts and stops.

Want to know more? Read the app brief, now available on our website.