Concordia Agents are mobile objects. During the lifetime of an agent, it can travel to different hosts on a network and perform work at each location. The mobility of agents differentiates them from either distributed objects or Java applets.
The purpose of this document is to describe the steps required for the development of Concordia Agents. Concordia is designed to hide the complexities of program mobility from users, so developing a Concordia Agent is not all that dramatically different from developing a non-mobile program.
This document describes the application programming interface of the Concordia system and details some issues that should be considered during the Agent development process.
The next section of this document introduces a simple agent application. Subsequent sections incorporate more advanced features of Concordia into the agent application. The topics covered by this document include: basic agent construction, collaboration, distributed events, access to native services, and the lightweight agent transporter.
On Windows systems, the source code for the examples in the document can be found in the directories under %CONCORDIAHOME%\Agents\examples\DeveloperGuide. On UNIX systems, the code is available at $CONCORDIAHOME/Agents/examples/DeveloperGuide.