Agents on the Web Making Agents Secure on the Semantic Web

A re you one of those people who never shops on the Web for fear of fraud? Does the thought of online banking make you antsy? Do you get nervous when talk of digital government arises? If so, you're not as paranoid as your friends say — some of your concerns are indeed well founded. With the rapid development of Web technologies and applications built on these technologies , new security risks have emerged. Traditional security models do not provide adequate protection in this dynamic and open environment. Fortunately, as we'll discuss, significant efforts are under way that should make Web services secure. Nowadays, an individual still might be able to avoid going online, but any company that wants to compete in the global economy needs a Web presence. That entails risks. To balance Internet access needs with security concerns, the business world needs technologies that provide security with high assurance. Security leaks can cost a company its competitive advantage, and leaks are rarely accidental. Industrial espionage has been around for ages. In the 16th century, for instance, Venetian merchants used poison and bribery to gain market share. As emerging Web services become preeminent in e-commerce, advanced technologies, such as software agents and ontologies, will allow more subtle attacks. Security is not solely a problem for human organizations. In a real scenario reminiscent of the cartoon Spy vs. Spy, an assassin agent recently connected itself to the JADE (Java Agent Development Framework) agent platform in Italy, killing all of the agents it found there. Although not as serious as it might appear — because the agents in Italy were there for research purposes — the episode nevertheless shows that agents need security. As agents adopt more mission-critical roles within enterprises, this concern will grow in importance. The World Wide Web was designed for humans. The envisioned Semantic Web is geared toward agents as well. Semantic constructs — such as ontologies represented in the DARPA Agent Markup Language (DAML), the Resource Description Framework (RDF), and XML — will let agents, as well as people, understand a Web page's content. The Semantic Web's success will depend on the implementation and use of Web services, which will likely be agent-based in the future. Using intelligent collaborations, agents can achieve global optimization while adapting to local requirements. 3 This approach will let agents use the large amount of information available over the Web — …