End-users can create and launch a system of agents to monitor diverse information sources (CHAINPlan™) such as relational data bases, web sites, sensors, and many others and alert them by email or phone messages (CHAINMail™) whenever user-defined critical situations are discovered.

The CHAINPlan™ can act as building blocks (CHAINLinks™), enabling users to create larger, more complex plans (CHAINClusters™) based on existing plans.

Global InfoTek, Inc (GITI) developed CHAIN as an intelligent multi-agent system that allows autonomous agents to be created, linked together and managed in order to provide extensive capabilities to users. CHAIN focuses on the ability to intelligently monitor information from diverse data sources, allow the user to provide specific logical constraints and rules to be applied to this data and specify who should be notified when specified conditions have been met.

CHAIN provides the ability to compose Plans for monitoring events, each Plan consisting of a set of agents linked together in a logical sequence to monitor and analyze information and notify interested parties. The user interface allows the user to see the set of available agents and graphically construct the Plan, indicating the sequence in which the agents will be executed and what data each agent will generate and/or receive from the other agents.

CHAIN allows the user to provide advanced scheduling constraints upon a Plan as well. A user is able to specify that a Plan should execute autonomously at a particular date and time in the future and and may instruct the Plan to restart on an hourly, daily or a user-defined interval. Hence, Plans may be executed without the user even logged into the system.

CHAIN is backed by the ISL system. This provides a powerful solution for rapidly integrating stove-piped information systems into agents that CHAIN may dynamically discover and incorporate into new and existing Plans. These agents may be classified into various sets of agents including, but not limited to: Retrieval, Analysis and Notification agents. Examples of these types of agents include:

• Retrieval Agents
• Web sites
• Web services
• Databases
• Sensors
• Applications (any data source)
• Analysis Agents
• Web site content analysis
• Numerical analysis
• Geographical analysis
• Notification Agents
• Email
• Phone / Pager
• Application
• Web page Generation
• Verona Notebook

Overall, CHAIN provides a powerful, dynamic, multi-agent system allowing the agents to be “chained” together to create complex ad-hoc behavior. Now, the processing of the individual tasks performed by these agents can be pooled together to autonomously accomplish a far greater and more complex task thereby freeing the human resources to work on other mission critical human-based tasks.