03.01.2015 Views

Combining Information from Multiple Internet Sources

Combining Information from Multiple Internet Sources

Combining Information from Multiple Internet Sources

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

2.1 Design of the application part<br />

The application part is written using the Java language and utilizes the agent platform – JADE<br />

which is also written using Java. The web application hence it requires the Java Servlet Container –<br />

for this purpose Apache-Tomcat was used.<br />

To create the testing tools the JADE Agent platform was utilized. Agents provide us with an<br />

abstract layer which can reflect the real-life situations. Agents play the role of experts at knowledge<br />

about given queries by hiding <strong>from</strong> the user the presence of the real search engines. Though the<br />

usage of software agents was not necessary it was an interesting and easy way to write an<br />

application; even such one which is written only to aid in testing something different than a multiagent<br />

system. Agents provide an easy interface of communication between objects and modules and<br />

that makes developing applications relatively easy. Though there is a variety of agent platforms like<br />

IBM Aglets or ZEUS the decision was made to use JADE since it is still actively developed whereas<br />

other agent frameworks are not. Summarizing, the tool could be written without any agent platform,<br />

but development of agent based environments is fairly easy and, at the same moment, very elegant.<br />

Next part presents the detailed design of the application. In general the application can be<br />

divided into two modules: Client and Main. The Client module is responsible for serving user<br />

requests and forwarding those to the web part of the application. Client module is responsible for<br />

interacting with the end-user; The Main module receives requests <strong>from</strong> the Client module and<br />

manages necessary agents for information retrieval and processing of the retrieved results. The<br />

Main module could also be accessed <strong>from</strong> any application (standalone/web based) as its entry point<br />

is started as a separate thread in the application.<br />

Next part of the chapter describes the Client module design details. Its information flow,<br />

workflow and functionality will be described. Afterwards, in the next part of the chapter, in similar<br />

manner, the Main module will be presented.<br />

6

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!