Collaborative Business Items (CoBIs)

Collaborative Business Items (CoBIs) is an EC funded project which bridges the gap between enterprise systems and networked embedded systems. Project partners are: SAP, Infineon, BP, TecO, Lancaster University, University of Twente and Ambient System.
Christian was the technical platform lead and represented TecO as the largest academic partner within the consortium.

Synopsis

CoBIs researches the integration of wireless sensor networks in business processes. Our approach is to bring parts of a business process to execution directly within sensor networks. Collaboration among the sensor nodes enables the execution of complex tasks and re-integrates the results in backend systems.

Application

The following figures depicts an example of a CoBIs scenario for asset management and workplace safety in industrial environments. CoBIs nodes can collaborativly detect hazardous situations for chemical containments, e.g. wrong storage area, exceeded storage limit, incompatible (inflammable) combination of chemical containments.
In the example below chemical drums are represented by dummies. The CoBIs nodes detect whether the storage limit is exceeded or not. Hereby, no backend interaction is necessary. The CoBIs nodes on the drum replacements execute business logic and decide collaboratively about hazardous situation.


Fig. Storage limit achieved (condition green)	Fig. Storage limit exceeded (condition red)

At TecO we are working on capabable wireless sensor nodes (Particle nodes) and a service-oriented architecture for integration. This approach results in an overarching concept since CoBIs employs not only one platform but many different. Apart from Particle Computer from TecO, the other platforms are Infineon's Sindrion platform and Ambient Systems µNodes. CoBIs can be applied in various scenarios. We especially investigate safety scenarios from BP involving storage of chemicals and worker safety. The approach can also be applied to electronic seals.
Finally, this CoBIs scenario was implemented in BP's chemical plant in Hull, UK. The following pictures illustrate the trial.


Fig. CoBIs installation	Fig. CoBIs Installation (close-up, CoBIs Particle Nodes)

Several trials were successfully conducted in 2006 and 2007.

Technology

Requirements for a CoBIs technology are

Sufficient computing power to execute business logic on the items
Wireless communication capabilities to enables collaboration between business items and to organise distributed processes.
Sensing capabilities for detecting item-to-item proximity and gathering environmental information
Remote update for service deployment, i.e. moving business logic on the items, update and service reorganisation.
Storage for services, but also for intermediate results, sensor data, tracking data, and logging

A possible instance of a CoBIs node is the DigiClip. The picture on the right depicts the technical characteristic of this node.


Fig. CoBIs Node (DigiClip) (click to enlarge)	Fig. CoBIs Node (CatPart) (click to enlarge)

A second generation, CatPart, of nodes was develop including more sensors and a infrared (IR) location system packaged in a more robost housing for industrial trials. A photo of the hardware is shown on the right picture above.

Results

CoBIs was successfully implemented for a period of several weeks in 2006 and 2007 in one of BP's chemical plant in Hull, UK.
By collaboratively updating the storage state of each service instance, storage regulations can be checked locally on each node. Local alarms are then also signalled collaboratively within a location. The hazardous situation due to storage violations was recognized within 1 second. According to BP up to 10 seconds are acceptable. In our experimental setting, flashing of the bright LEDs indicated the hazards. Clearly, the advantage is in-situ, real-time reaction by shortening the communication path and let the items directly reason on the situation without an permanent connection to a supervisionary backend system.
Relocated business logic utilizing collaboration among the sensor nodes keeps information local within the sensor network. Collaboration reduces the message load in case of hazardous detection events from 10 messages/s down to 2 messages/s. The complete history could be downloaded for instance during regular check or maintenance intervals.

Collaborative Business Items (CoBIs)

Synopsis

Application

Technology

Results

Further Readings