7.4 Automated Web Services

7.4.1 Introduction

When looking towards the future of Web services, the breakthrough will probably come in the form of access to services and added intelligent capabilities through semantic reasoning. Agents bring the most crucial capability to turn Web services from the existing mass of information, where users need to surf and browse, into a dynamic set of capabilities deployed around and serving the user . Agents represent this great opportunity towards a new and completely different computing model, freeing humans from many chores imposed by the Internet.

Agents can turn Web services into proactive entities working as peers to serve the end users, representing them and defending their interests in a competitive world where services are negotiated and composed dynamically. Agents introduce an unparalleled level of autonomy into future systems so that users can delegate high-level tasks in a generic manner. To enable agents, they need to get mission statements, the definition of domains of competence, and a definition of autonomy through policies to be applied in these domains.

These domains and policies need to support reasoning as they might overlap in real life; for example, the policies of multi-national companies have to comply with the policies of the nations in which these companies operate . These domain and policy mechanisms permit deployment and dynamic adaptation to any situation. In particular, they allow Web services to combine without prior design, to negotiate end-to-end contracts to ensure the final result of composed Web services, and to monitor their execution on behalf of the user.

Agents will become the trusted intelligent interface between man and machine, allowing communications through speech acts and representing the interest of the user in any Web transaction at any time, like a trusted friend or broker. Hence, agent interfaces need to evolve towards ease of use, ease of delegation and monitoring of tasks, increased privacy, personalization and security, and user habits being acquired through learning.

Agents can now migrate to slim wireless appliances and evolve in a multitude of micro- worlds , typically the cells of wireless phones, malls, schools , and a community of friends ; they can discover the resources and represent their user. This technology needs to be rolled out on a large scale to test the deployment capabilities as well as the usability of the technology in a mass market.

7.4.2 Machine-to-Machine Work

We've focused thus far on how components or agents can accept delegation, perform tasks on our behalf, understand what we want to do, and understand how to do things effectively to change the nature of the relationship between us and machines. But, equally profoundly, they change the nature of the way machines interact with machines. Although some degree of automation and humanless interaction occurs today, we expect to see an explosive increase of this type of activity as millions of people delegate billions of tasks to agents.

Each agent may rely on many others to do its work, and these in turn will rely on others and interact with others. This "Web" of machine-to-machine work will change the landscape of the computing infrastructure and the nature of transactions. This can provide a new level of reality that begins to approach the visions of science fiction writers. Once the machines are empowered and capable of doing real work on their own, what governance will they need and what constraints will be imposed?

Many authors have spent a considerable amount of time on how the world would change if everything is done automatically by machines or robots. Not only science fiction authors but also futurists look at this topic with a mixture of fear and desire . George B. Dyson, for example, provides visionary , long-range concepts in his book ^[2] and Forrester does some research on machine-to-machine work, which they call the X Internet. ^[3]

^[2] George B. Dyson (1998). Darwin among the Machines: The Evolution of Global Intelligence . Cambridge, MA: Perseus Publishing.

^[3] http://www.forrester.com/ER/Marketing/0,1503,214,00.html

The essential points of these studies are that the Internet provides a foundation for an almost incomprehensible explosion of machine-to-machine interactions, fundamentally altering our concept of commerce and vastly expanding the range of value-added transformations that constitute the basis of our economy.

7.4.3 The Semantic Web

The easy information access based on the success of the Web has made it increasingly difficult to find, present, and maintain the information required by a wide variety of users. In response to this problem, many new research initiatives and commercial enterprises have been set up to enrich available information with machine understandable semantics. Only this enables automated machine-to-machine transactions. This semantic Web approach, which was defined by Tim Berners-Lee, provides intelligent access to heterogeneous, distributed information, enabling agents to mediate between user needs and the information sources available.

Web services deal with the limitations of the current Web. Currently, the Web is mainly a collection of information but does not yet provide support in processing this information. Web services can be accessed and executed via the Web. However, these service descriptions are based on natural language descriptions (e.g. "Find the cheapest flight to London"). To interpret these descriptions, a human programmer needs to be kept in the loop. As a result, the scalability and economic potential of Web services is severely limited. Reaching the full potential requires utilizing semantic Web technology as a basis for automating access to Web services. Machine-interpretable semantics will enable us to mechanize service identification, configuration, comparison, and combination.

Semantic Web-enabled Web services have the potential to change our life to a much higher degree than the current Web already has done. The following elements have been identified to be the minimum requirement for a well-working semantic Web: public process description and advertisement; discovery of services; selection of services; composition of services; and delivery, monitoring, and contract negotiation. These elements were described in Chapter 6 in more detail. Here, you can see how they are related to agents.