Sukun Kim Rufus McLain MBA 290/ENG 298A – Introduction to the MOTMBA 290/ENG 298 – Introduction to MOT Sukun Kim, Rufus McLain Page 1 of 19 Wireless Sensor Networks – An Introduction Advances in semiconductor technology following Moor’s law have enabled enormous growth in information technology: PC, Internet, wireless phone, etc. Technology is now passing a point where a battery-powered computer is powerful enough to operate wireless communication devices for an extended period of time. As a result, wireless sensor networks (WSN) have been developed in universities and research labs, paving the way for companies like Dust Networks. The following paper looks at the market for WSN and analyzes ways in which Dust Networks will be able to thrive in this new and growing environment. Dust Networks & Wireless Sensor Networks Founded in 2002 by Kris Pister and Rob Conant, Dust Networks provides low-powered wireless sensor networks to solution providers, integrators, and OEMs for monitoring and controlling applications. (See Exhibit 1 for a history of the Dust Networks) A wireless sensor network is a wireless network of small computers with sensors and actuators. These small computers are called motes and they are composed of three principle parts: a senor or actuator, a microprocessor, and a radio chip. The sensor reads physical data of the real world, the actuator can control basic functions like turning off a light or changing a room temperature, the microcontroller processes data from the sensor and makes decisions, and the radio chip communicates data wirelessly. The next element of the network is the SmartManager that collects all the information from the motes and acts as a gateway betweenMBA 290/ENG 298 – Introduction to MOT Sukun Kim, Rufus McLain Page 2 of 19 them and the server. Throughout the network and on the server there is software that executes routing, timing, network optimization and management functions to ensure that connectivity and long-term reliability. (See Exhibit 2 for a screen shot of the software) History of WSN Technology Wireless Sensor Networks started as a DARPA (The Defense Advanced Research Projects Agency) sponsored project named NEST at UC Berkeley in 2000. NEST (Network Embedded Systems Technology) is used for unmanned monitoring of enemy battlefields. Unmanned air vehicles would drop motes that would use magnetometers and sounder to detect enemy movement. These networks would be able to determine the type, speed, and direction of ground vehicles. Soon after the success of this project, Intel saw the potential of this technology and hired, then DARPA manager, David Tennenhouse as a research manager. Intel then funded a company called Crossbow to work with UC Berkeley to further develop WSN technology. In 2002, Kris took leave from the research group at UC Berkeley to start Dust Networks. Since this time there have been many successful implementations of Wireless Sensor Networks. Practical Applications of the Technology Machine Monitoring Imagine a motorcycle company and that uses manufacturing plant with a conveyor system and machines with bearings which wear out over time. When the bearings completely wear out or fail, the machine stops and the entire conveyor stops as well. It takes between two and three hours to repair the system and the entire downtime costs hundreds of thousands of dollars. However, by monitoring the vibrations of the machines, the company could determine whether a bearing is almost worn out and schedule the bearing replacement in advance during a routine downtime. One possible way to monitor vibration is by attaching accelerometers to machines andMBA 290/ENG 298 – Introduction to MOT Sukun Kim, Rufus McLain Page 3 of 19 then wiring them to a central control room. However, wiring one accelerometer costs around one thousand dollars and if hundreds of machines need to be monitored, this cost becomes very high. A wireless sensor network could be used to solve this problem at a fraction of the cost. Seismic Testing ($200 vs. $5,000 per node) In the department of civil and environmental engineering at UC Berkeley, Professor Steve Glaser performed an experiment to look at the effect of an earthquake on a structure. A three-story model of building was excited on a shaking table to simulate an earthquake. Vibrations measured before, during, and after an earthquake could be used to determine the structural integrity of the building and pinpoint where the building was damaged. Before using wireless networks, the experiment to much longer because the time needed for wiring and the cost per node was $5,000, compared with $200 for wireless nodes. HVAC Energy Testing Nodes ($100 vs. $500 per node) Dust Networks partnered with Honeywell on a project to install wireless sensor networks in a chain of grocery stores. These networks monitored electricity usage from lighting and temperature controls in an effort to analyze ways to save money. The wireless motes cost $100 verses the $500 per sensing point with a conventional system. The installation was also done in only three hours, versus three to four days. Market Size According to the Wireless Research Group, the market size for wireless sensor networks will be approximately $5billion this year and over $8billion in 2007. The number of units is expected to grow to just over 700 million by 2007, catching up with the number of wireless handsets. Building and industrial automation, two of Dust Networks’ primary application foci, will constitute more than have of the total market.MBA 290/ENG 298 – Introduction to MOT Sukun Kim, Rufus McLain Page 4 of 19 Value in System Integration and Support Services (Module 4) There are three layers in the value chain for the wireless sensor network market; component manufacturing, network manufacturing, and system integration. There are established players in each of these levels and different potentials for current and future economic value. In order for Dust Networks to capture the largest amount of this value as possible, they must correctly position themselves within this value chain. (See Exhibit 3 for a diagram of the value chain and the major players) Component Manufacturing The first layer of the value chain is comprised of companies that build the components that