Companies of all sizes are deploying sensor technologies to track assets and automatically detect changes in temperature, motion, location, and other physical conditions. But it is only when sensor data are fully integrated with enterprise applications—such as inventory management, warehousing, and manufacturing systems—that the information allows organizations to gain a more complete understanding of what is happening in their physical environment. Sensor-based services now allow organizations to capture, manage, analyze, and respond to data collected from a wide range of sources.
Integration in Action
An example of this integration can be found at Dryden Flight Research Center, NASA's primary facility for conducting atmospheric flight research. Dryden uses a variety of precision sensors to obtain accurate flight test data. With help from Oracle, the center's ChemSecure project has integrated radio frequency identification (RFID) and sensor-based technology with its existing back-end applications to better manage hazardous materials, enhance security, and reduce supply chain costs.
Located at Edwards Air Force Base in southern California's high desert, Dryden manages many hazardous chemicals used in the maintenance of aircraft and other vehicles. These chemicals are stored in specially designed on-site sheds. For reasons of safety and efficiency, Dryden has equipped the chemical containers with RFID tags, installed temperature sensors in the sheds, and issued RFID-enabled ID badges to every employee authorized to handle chemicals. These devices are connected to the organization's back-end Hazardous Materials Management System (HMMS).
The HMMS houses complete records of all chemicals on the premises, including purchase dates, physical properties, and specified storage temperatures. The HMMS also contains information about which chemicals each employee is authorized to access, based on individual certifications.
When employees enter a shed to pick up a particular chemical, their ID badges are scanned. The HMMS makes the necessary associations and correlations to determine if the employee is taking out the appropriate chemical. If the wrong chemical is picked up, a yellow warning light comes on, and the employee is given a grace period to replace it. If the employee does not react properly, an alert is sent to management.
Because Dryden is located in the desert, heat is an issue for temperature-sensitive chemicals. If the temperature in a shed exceeds a set threshold, the sensor, working with the HMMS, triggers an alert, calling in the necessary personnel to correct the situation.
By linking sensors with back-end applications, the NASA field center can better capture real-time information about its assets. Dryden's ChemSecure project uses Oracle Database and Oracle Fusion Middleware to capture data from multiple sensors and integrate the RFID/sensor events with applications and systems.
Dryden has taken the project a step further by creating new operational processes. For example, ChemSecure includes a mobile component that allows a person at the scene of a chemical spill to use a handheld reader to scan details about the spill and send alerts to first responders (see Figure 1). The first responders, also equipped with handhelds, receive immediate instructions on how to resolve the issue, including information such as the substances they can safely use to clean up the spill.
Figure 1. Custom temperature sensors and stationary and handheld RFID devices work together to provide critical event and application information on the handling, storage, and access authorization of hazardous materials at NASA's Dryden Flight Research Center. Real-time event alerts are enabled via Oracle UI's and wireless Web and short message service alerts to operations and security groups.
Integrating Sensor Data
When it comes to integrating sensor data with enterprise systems, most organizations are not nearly so advanced as Dryden. For the most part, they have to learn to walk before they can run. This process typically occurs in three phases.
In the first phase, the organization is simply trying to capture data from a variety of sensors. But the information cannot be pieced together into one coherent picture because sensor data come in a variety of formats. Moreover, the data are often stored in information silos, making it difficult to interpret events. Therefore, the first priority is to integrate the sensor data.
An organization in this phase can use the Oracle Sensor Edge Server, a component of Oracle Fusion Middleware and Oracle sensor-based services, to bridge the gap between the physical world of sensors and the IT world of enterprise software systems. Designed to handle rapidly changing sensor standards and capabilities, the server shields the application developer from protocol and hardware changes and variations between different devices, so switching vendors or upgrading equipment can be done without affecting the implementation of the application.
The server furnishes a plug-and-play driver framework that lets you easily integrate data from hardware (e.g., RFID readers and printers, response devices, and sensors) with existing or new applications. The driver framework provides a standardized interface for the hardware. When a device connects with the server through the driver framework, all communications are in the same format. The driver framework converts sensor-event data into the underlying hardware protocols implemented by the particular driver and manages the communications with the device. The server also manages and monitors the health of the sensors.
Once the sensor data are captured, they are filtered, converted to a standardized format, and stored in the sensor data repository. Rules representing the conditions that you are interested in are continuously evaluated as the sensor data evolves in the repository. These rules interpret the data to identify events based on the evolution of many sensor measurements over time, providing a means to recognize even the most complex operational events. These events can then be distributed to any RFID application, service, or data store. Because the repository archives all measurements and observations, it provides the means to audit, troubleshoot, and query for important information not originally deemed relevant, and mine for valuable insights that optimize operational processes.
The second phase entails delivering sensor data to existing enterprise applications and thereby extending processes. For example, let's look at a company that makes ice cream. After the ice cream is produced and packaged, it's loaded onto a delivery truck for shipment to the customer. What happens if the truck's refrigeration system suddenly goes on the blink? In a typical scenario, a temperature sensor on the dashboard may alert the driver, who will then have to contact the service center and give the truck's exact location and identify the order that is being shipped and the customer it is going to. In some cases, the driver may not even have the answers to these questions.
But if the data are automatically captured and fed into the back-end applications, the system can trigger the appropriate responses when the temperature sensor data are received. Using Oracle Fusion Middleware, you can automate processes to immediately issue a service order and notify the manufacturing application about the spoiled delivery. Business-process modeling tools based on standards such as Business Process Execution Language for Web services let you assemble a set of discrete services into an end-to-end process, integrating not only sensor data but also data from your enterprise systems. Business Activity Monitoring capabilities enable you to detect, correlate, and analyze separate events, such as sensor readings. With Oracle Application Server Wireless, alerts can be sent to company personnel via e-mail, short message service, multimedia message service fax, or voice alerts in real time to keep them up-to-date on the activities.
Evolving the Business
The third phase involves not just linking sensor data with your internal applications but also with those of your business partners—and the world at large—to gain total situational awareness. This is where companies can attain actionable operational insight by having visibility into information across the entire organization. The ultimate goal is to coordinate business processes among trading partners through global information sharing. Real-time visibility into trading-partner data is a powerful notion.
Here's one example. The maker of tennis racquets places RFID tags on the finished goods at the item level to enable identification based on a unique ID. Next, the tennis racquets are shipped to the manufacturer's warehouse, where RFID tags at the case and pallet level enable hierarchical tracking and verification. The manufacturer's warehouse then ships the racquets to the retailer's distribution center. The RFID tags enable the manufacturer to record shipment time, contents, and location, while the retailer can verify shipment authenticity and record shipment receipt. The retailer tracks product availability and inventory from the distribution center and verifies product authenticity. Finally, inventory is updated in real time as products are sold.
Avoiding Integration Pitfalls
It's important to remember that as organizations progress from one phase to the next, the integration process must continually evolve with the overall business. That's why it's essential to have a flexible service-oriented architecture (SOA) that allows organizations to adapt to changes and accommodate new processes engendered by their sensor deployments.
Some companies struggle as they try to move beyond a basic sensor deployment and attempt to integrate sensor data with their back-end enterprise systems. The result can be an integration nightmare as they try to bolt sensors that have been deployed as stand-alone systems into their existing enterprise architecture.
Oracle addresses this problem with Oracle Fusion Middleware, a collection of standards-based software products that support the development and management of an SOA. The middleware is hot-pluggable, allowing organizations to leverage their existing investments in sensor technologies while avoiding an architectural mess.