|
|||||||||||||||||||||||||
|
|||||||||||||||||||||||||
|
The Information Personalization Agent Manager (iPAM) is a software framework that facilitates building information channels. iPAM serves as a "glue" layer between information suppliers and information dissemination tools. iPAM consists of information agents and a manager of those agents, built as a single Java "executable." iPAM is an information administrators toolset that allows administrators to build information collection and dissemination services. Such services, or "channels" can be based on:
iPAM is "glueware" that can be used to bridge information sources to sinks with dissimilar formats and/or protocols. iPAM has an open architecture that allows administrators to write custom agents to get, process, or distribute information. Small custom Java agents can be added to the iPAM toolset and easily interconnected with existing ones. iPAM is not an end user tool. It does not address content browsing, or subscribing for content. Many commercial tools by many vendors address those needs. Example channels:
|
||||||||||||||||||||||||
The software is written in Java. The architecture makes extensive use of Java Interfaces. There is a built in web service as well as support for mail, ftp, and many other protocols. Further information can be found in the developer's documentation.
|
|||||||||||||||||||||||||
The heart of iPAM is the executive. It provides the services that Inputs, Outputs, Handlers, and Daemons use. Some of these services are tracking directories, understanding the mapping between working directories and remote URLs, and scheduling. The executive can be configured via a number of parameters.
|
|||||||||||||||||||||||||
Input agents are the entry point for information into the system. They are responsible for getting information using the appropriate protocol and converting the input into a standard internal representation.
|
|||||||||||||||||||||||||
Handlers are agents that handle the information on its way from the sources to the outputs. They can be activated on a time schedule to collect/aggregate products into information channels. Jugglers can accept input from source agents sources and provide output to multiple output agents. Handlers share several common configuration variables.
|
|||||||||||||||||||||||||
Output agents can distribute products. They can be connected to multiple Sources and/or Handlers simultaneously.
|
|||||||||||||||||||||||||
Daemons are agents that can be scheduled (just like a Handlers) but they have neither inputs or outputs. See schedule for information on scheduling options.
|
|||||||||||||||||||||||||
Test |
These agents are useful for testing and for making demonstrations. Most of then input agents create different types of "canned" products.
|
||||||||||||||||||||||||
iPAM can output via email, BackWeb, Castanet and StarBurst MFTP or it can display information on Object/FX SpatialX maps. There is also message input and output using America Online's Instant Messenger service. iPAM is extensible with custom Java "plugins." Contact cda@mitre.org for further information. |
|||||||||||||||||||||||||
Revised: 8 March 2000 |